Defining “recklessness” in research misconduct proceedings

ABSTRACT

To find research misconduct in research that has been supported by federal funds, an institution must determine that the misconduct was committed intentionally, knowingly, or recklessly. “Intentional” and “knowing” are straightforward standards. Yet “reckless” often mystifies institutions, which struggle to assess whether a respondent’s conduct should be deemed “reckless,” or merely negligent. This difficulty is most pronounced when allegations are lodged against the author under whose supervision the primary research was conducted – most often, the senior and/or corresponding author of a published paper who may not have been directly involved in performing the experiments or preparing the data under scrutiny. In these situations, investigation committees and the institutional “deciding official” must assess whether the supervising scientist is guilty of research misconduct – based on the theory that their supervision of the research and development of the publication containing falsified, fabricated, or plagiarized information was reckless – even if that person did not perform the experiment or assemble the research records in question. This paper seeks to provide a framework for evaluating the circumstances in which past supervisory conduct should be deemed “reckless” and thus a basis on which a finding of research misconduct may be made.

KEYWORDS:

• Research misconduct
• recklessness
• ORI
• NIH; responsible conduct of research

Introduction

Since 2005, when the concept of “recklessness” was incorporated into the research misconduct regulations at 42 C.F.R. Part 93 (“Part 93”) governing work supported by U.S. Public Health Service (“PHS”) funding, institutions have grappled with the standards needed to find research misconduct predicated upon recklessness of a respondent. Part 93 states that a finding of “research misconduct” requires demonstration by a preponderance of the evidence that “falsification, fabrication, or plagiarism” was “committed intentionally, knowingly, or recklessly.” While intentional, knowing, or reckless falsification, fabrication, or plagiarism constitutes research misconduct, “honest error” by an author that results in falsification, fabrication, or plagiarism of material is not considered to be research misconduct.

The “intentionally” and “knowingly” standards are readily understood with reference to the plain meanings of such terms in everyday use. “Intentionally” means that the research was carried out with the respondent’s specific intent to falsify, fabricate, or plagiarize. “Knowingly” equates to a lower standard of culpability than “intentionally” because, while it implies the respondent knew the research misconduct was carried out, it does not require the respondent to have intended the research misconduct to have been carried out (for example, having no direct involvement in the misrepresentation of results but knowing that certain results presented by a coauthor were not consistent with earlier iterations of those data presented by a coauthor).

Unlike knowing and intentional conduct, however, reckless conduct cannot be easily defined with reference to an everyday standard, nor is “recklessly” defined under Part 93. Participants in research misconduct proceedings typically understand recklessness to fall somewhere between, on the one hand, intentional or knowing conduct (both of which constitute research misconduct) and, on the other hand, conduct that is simply negligent. Negligence, like honest error, is regarded as a state of mind devoid of any intention to deceive and is widely understood to differ from behavior meeting the definition of research misconduct (Committee on Science, Engineering, and Public Policy 2009). However, a broad range of conduct can fit between “intentional” acts to fabricate, falsify, or plagiarize, and “negligent” conduct that results in fabrication, falsification, or plagiarism. Institutions and their faculty research misconduct committees often struggle to articulate and apply a satisfactory and consistent recklessness standard to allegations of research misconduct.

In our experience, and as demonstrated in recent high-profile cases involving prominent scientists and/or institutional leaders as respondents, this issue arises most often in research misconduct proceedings concerning respondents who have supervised, but not themselves performed, the research at issue. Often, the supervisor in such circumstances is the director of the laboratory staff or research team which has conducted the relevant research; this director frequently assumes the role of senior and/or corresponding author of a publication that has become subject to allegations of falsification, fabrication, or plagiarism.¹ In such cases, if data appear to lack integrity, it is very common for the senior scientist or research leader to take the position that they (1) did not know about the data integrity problems and (2) could not reasonably have identified or anticipated those problems. The issue thus becomes, given the respondent’s supervisory responsibilities over the work in question and status as senior and/or corresponding author, whether the respondent’s failure to identify the problems resulted from recklessness in their supervision of the research team or whether the respondent’s conduct falls short of recklessness, even if the committee has identified supervisory shortcomings that may have contributed – even directly – to the falsification, fabrication, or plagiarism of data.

Inquiry and investigation committees, most often constituted by unconflicted peer faculty members of the respondent, face challenges in assessing this issue. Committees typically comprise experienced researchers, selected by their institutions because of their relevant subject matter expertise and reputation for academic excellence, research integrity, and independence. Many committee members have significant insight in evaluating what oversight practices should have been applied by the respondent in light of the type of research being performed and the respondent’s overall responsibilities. Nevertheless, the vast majority of investigation committee members lack forensic experience in assessing the conduct of their peers according to the unique regulatory framework prescribed under Part 93, which requires them to apply difficult concepts like recklessness, “significant departure from accepted practices in the relevant research community,” “preponderance of the evidence,” and burden-shifting to the respondent to demonstrate “honest error.” Their lack of experience in applying these types of legal standards becomes particularly challenging when committee members are not provided with clear guidance as to conduct that would or would not constitute recklessness. Further, determinations by these committees must be considered and approved or rejected by each institution’s “deciding official,” most often the chief academic officer or vice provost for research, who, like the committee members, is typically also confounded by the application of the recklessness standard.

This article analyzes the recklessness standard in research misconduct. There are presently few tools available to determine whether a respondent’s conduct constitutes recklessness and scant legal analysis of the standard in the research misconduct context. These deficiencies contribute to the difficulty that inquiry and investigation committee members, and deciding officials, experience when seeking to make findings in a research misconduct case. Our discussion and analysis proceed as follows: First, we provide background on the recklessness standard under Part 93 and related research misconduct regulations and policies and describe how that standard has been applied in the context of both research misconduct proceedings and other areas of law. Drawing on this information, we then set forth a clear standard for recklessness that can be used by investigation committees when determining if a respondent should be found to have committed research misconduct.

We conclude by offering some additional commentary on the possible policy implications related to the definition of recklessness provided in this article. In short, we think that additional guidance detailing how the recklessness standard should be applied to various fact patterns involving the work of a supervising scientist is needed from either federal agencies with oversight over federally funded research or professional societies. ORI would be particularly well-positioned to provide the research community with additional guidance, including a definition of “recklessly,” and/or case studies detailing how research institutions or ORI have applied the recklessness standard to various research misconduct cases. Such guidance on the recklessness standard could also be used to enhance existing materials and develop additional trainings regarding the role of the senior author and others with oversight responsibilities in a research environment. Currently, the literature on best practices for research supervision, training, and mentorship is not robust, even though these topics are referenced by the National Institutes of Health (“NIH”) as key topics to include in Responsible Conduct of Research (“RCR”) training (National Institutes of Health 2022). In our view, clarification of the recklessness standard for research misconduct proceedings and examples from ORI and/or professional societies as to how the standard has been correctly applied in various research misconduct proceedings could help inform the broader discussion regarding fundamental standards for laboratory supervision and training at research institutions across the United States.

Brief history of recklessness as a regulatory standard

The first iteration of the research misconduct regulations applicable to PHS-funded work were issued by PHS in 1989 (U.S. Public Health Service 1989, 32447). The 1989 regulations defined “misconduct in science” at 42 C.F.R. § 50.102 as:

[F]abrication, falsification, plagiarism, or other practices that seriously deviate from those that are commonly accepted within the scientific community for proposing, conducting, or reporting research. It does not include honest error or honest differences in interpretations or judgments of data.

These regulations, which were in effect until 2005, thus did not specify any state of mind needed to reach a finding of misconduct. However, in a 1993 decision by the U.S. Department of Health and Human Services (“HHS”) Departmental Appeals Board regarding ORI findings against a respondent, the HHS administrative law judge (“ALJ”) explained that there was a distinction between honest error and statements that are made “with knowledge that they would mislead the reader” (Sharma). This portion of the (1993) ALJ decision was later cited as evidence that ORI and its predecessors have always applied an intent standard, even under the pre-2005 regulations (Brodie 2010).

On December 6, 2000, the White House Office of Science and Technology Policy (“OSTP”) issued a government-wide Federal Research Misconduct Policy. This policy “applies to federally funded research and proposals submitted to Federal agencies for research funding” and required that all “Federal agencies that conduct or support research … implement this policy” within one year of the publication of the policy (Office of Science and Technology and Policy 2000, 76262). Like the standards later adopted under Part 93, the Federal Research Misconduct Policy provides that a finding of misconduct requires that “[t]he misconduct be committed intentionally, or knowingly, or recklessly” (76263).

In the preamble to the Federal Research Misconduct Policy, OSTP acknowledged that several commenters had requested clarification regarding the level of intent that is required to be shown in order to reach a finding of research misconduct (76260). However, OSTP did not proceed to define “intentionally, knowingly, or recklessly” or provide guidance that could help institutions determine how to evaluate cases in which recklessness would be at issue.

In response to the Federal Research Misconduct Policy, ORI revised its existing research misconduct policies, publishing a final rule in 2005 that is now codified at 42 C.F.R. Part 93. ORI adapted this same three-part standard for research misconduct in the 2005 regulations at 42 C.F.R. § 93.104. Similar to OSTP, ORI, when issuing the 2005 final rule, did not discuss the three mens rea standards.

In 2004, ORI published its proposed revisions to its research misconduct regulations. A key provision in the 2004 proposed rule was a rebuttable presumption of research misconduct “in the absence of, or respondent’s failure to provide, research records adequately documenting the questioned research” (HHS 2004, 20790). This rebuttable presumption would have created, in effect, a regulatory mechanism for concluding that failure to maintain records constitutes, at a minimum, recklessness by a respondent, absent a reasonable explanation for the absence of records. However, after commenters expressed concern for holding a respondent responsible for the retention of records over which they may have no control and a lack of guidance as to a “reasonable explanation” for the absence of records, ORI revised the regulation at 42 C.F.R. § 93.106(b)(1) to state:

The destruction, absence of, or respondent’s failure to provide research records adequately documenting the questioned research is evidence of research misconduct where the institution … establishes by a preponderance of the evidence that the respondent intentionally, knowingly, or recklessly had research records and destroyed them, had the opportunity to maintain the records but did not do so, or maintained the records and failed to produce them in a timely manner and that the respondent’s conduct constitutes a significant departure from accepted practices of the relevant research community.

Thus, the current operative federal regulations provide that the absence of original records should be viewed as evidence of research misconduct if the respondent’s failure to produce the records is a significant departure from accepted practices. This standard is particularly relevant when evaluating allegations of research misconduct against senior scientists or others who serve in a supervisory capacity. These individuals may have had no direct knowledge of the falsification, fabrication, or plagiarism, but their failure to retain original data or ensure adequate processes in the laboratory regarding retention of original data may itself – in some instances – be viewed as evidence of research misconduct and may lead to a finding of recklessness against that individual.

Application by ORI of the recklessness standard

ORI has never published official guidance on reckless intent. However, ORI has undoubtedly applied the recklessness standard in many cases in which the agency has conducted its own review of research misconduct allegations as part of its authority under 42 C.F.R. § 93.400 et seq. While ORI’s own written analyses and internal working documents are largely shielded from public review, ORI’s position on recklessness can be deduced in part from publicly available information regarding ORI’s findings of research misconduct levied against respondents, including in: (1) ORI notices of findings of research misconduct, which are published in the Federal Register; (2) administrative appeals by respondents of findings of research misconduct made by ORI/HHS, which result in publicly available ALJ opinions; and (3) filings made and opinions rendered during federal district court cases involving ORI research misconduct cases.

ORI case summaries

When ORI reaches findings of research misconduct against a respondent, a notice of findings is published as a case summary in the Federal Register. These case summaries typically do not provide significant insight into the standard for adjudicating recklessness, as they are presented at a high level, often using the three categories of mens rea interchangeably. A typical formulation is: “ORI found that Respondent engaged in research misconduct by intentionally, knowingly, or recklessly falsifying and/or fabricating data.”

As of July 29, 2023, there are 33 case summaries currently posted on the ORI website, representing cases in which administrative actions imposed by ORI are ongoing (Office of Research Integrity n.d..).² Of the 33 posted summaries, there are only two cases that cite recklessness as the sole reason for a finding of research misconduct:

• In a 2020 case summary, ORI found that Prasadarao Nemani, Ph.D., “engaged in research misconduct by recklessly including falsified and/or fabricated data in … published paper and grant applications submitted for PHS funds … [Dr. Nemani] recklessly reported falsified and/or fabricated image data for enterobacterial infection-induced intestinal epithelial cell injury in a neonatal murine model to falsely represent results using images from unrelated experiments in eight (8) figures included in one (1) published paper and four (4) grant applications” (Office of the Secretary, HHS 2020, 44911). This summary does not provide any additional detail or context regarding the recklessness finding.

• In a 2018 case summary, ORI found that Christian Kreipke, Ph.D., “recklessly caused or permitted twenty-three (23) instances of research misconduct in his three (3) grant applications, two (2) articles on which he was the first listed author, and two (2) posters on which he was the first listed author” (Office of the Secretary, HHS 2018, 38316). As with the 2020 case summary noted above, this case summary also does not provide any detail or context regarding the recklessness finding. However, the respondent, Dr. Kreipke, appealed ORI’s findings of misconduct to an ALJ of the Departmental Appeals Board (Kreipke 2018). As discussed below, the resulting ALJ opinion provides some guidance on the definition of reckless, and ORI has pointed favorably to that guidance (e.g., Office of Research Integrity 2018).

For the vast majority of the remaining 31 ORI case summaries, it is unclear, upon reading the summary, if research misconduct was determined by ORI to be intentional, knowing, or reckless.

ALJ and federal district court discussion of recklessness

In cases in which ORI reaches findings of misconduct under 42 C.F.R. § 93.400 et seq., respondents are provided with the right to appeal ORI’s decision to an ALJ. The ALJ decisions are publicly available, and several decisions provide insight and guidance into the recklessness standard under Part 93. In this section, we discuss four research misconduct cases that were subject to the administrative appeals process and resulted in one or more published opinions that include judicial commentary relating to the recklessness standard.

ORI v. Bois

In ORI v. Bois, ORI made findings of misconduct against Dr. Bois for failing to report results, publishing false data, excluding a control, failing to repeat experiments sufficiently to verify data, and altering figures (Office of the Secretary, HHS 2013, 23255). Dr. Bois appealed these findings.

In a 2011 decision, the ALJ denied Dr. Bois’s appeal. In her decision, the ALJ did not dive into great detail on the applicable standard for recklessness but did comment that respondent’s “failure to review all of his lab notes before reporting the results was certainly reckless” (Bois 2011, 17). The ALJ’s opinion ultimately did not turn on an assessment of recklessness, as the ALJ concluded that “the undisputed evidence establishes that at least some, if not all, of Respondent’s misconduct was knowing and intentional” (12).

Dr. Bois then brought suit against HHS in federal district court, alleging that the ALJ’s decision was “arbitrary, capricious, an abuse of discretion … in violation of the [Administrative Procedure Act] … [and in violation of] Dr. Bois’s Fifth Amendment right to due process” (Bois v US Department of Health and Human Services 2012, 3). The court remanded the case to HHS on these procedural grounds, but also found that “it was certainly reasonable for the ALJ to conclude, based on the undisputed facts, that ORI met its burden to establish – at the very least – reckless research misconduct, by a preponderance of the evidence” (5). The court emphasized:

And, when the mistake of omission (failing to report the contrary results from February) is combined with the mistake of commission (reporting on the results of the December experiment when that experiment lacked a control and did not in fact produce the reported results), the Court cannot find that it was arbitrary or unreasonable for the ALJ to conclude that ORI overcame the rather thin honest mistake defense by a preponderance of the evidence, and that Dr. Bois’s conduct was at least reckless. (Bois v US Department of Health and Human Services 2012, 8)

While additional appeals were pending, Dr. Bois and HHS reached a settlement (Office of Research Integrity 2013). The ALJ and federal district court decisions in Bois set forth the notion that the failure to follow the standard practice of verifying data prior to submission can be considered reckless. Notably, in the later-published notice of findings of misconduct against Dr. Bois, ORI did not describe “reckless” conduct, instead stating, “In the Settlement Agreement, the parties agreed that ORI found by a preponderance of the evidence that the Respondent committed misconduct in science and research misconduct by … [k]nowingly and intentionally falsely reporting … and [f]alsifying data” (Office of the Secretary, HHS 2013, 23255).

ORI v. Sauer

In ORI v. Sauer, ORI made findings of research misconduct against Frank Sauer, Ph.D., finding that Dr. Sauer “engaged in research misconduct by intentionally, knowingly, or recklessly falsifying and/or fabricating images in seven (7) submitted NIH grant applications and three (3) published papers by manipulating, reusing, and falsely labeling images” (Office of the Secretary, HHS 2017, 31334). Dr. Sauer appealed these findings.

In a 2017 decision, the ALJ denied Dr. Sauer’s appeal. The ALJ decision is significant because it rejected the respondent’s argument that he could not be found guilty of research misconduct if he did not generate the data himself (Resnik et al. 2017). Instead, the ALJ reasoned that the Part 93 definitions of fabrication and falsification “clearly contemplate the reporting of research results, even if the individual reporting the results did not create the data” and therefore that since the respondent “ultimately reported that data,” the reporter thereby adopted “the data, and its flaws, as his own” (Sauer 2017, 14). The ALJ continued that “even if saboteurs [had manipulated the data, as Dr. Sauer alleged], such actions … do not shield Respondent from research misconduct, as he did not ensure that his papers and grant applications were without fabrication and falsification when he submitted them” (15).

After laying out this view of the salient Part 93 definitions, the ALJ provided a framework for defining and assessing the recklessness issue:

Over and over, on numerous occasions, Respondent submitted grant applications and publications that contained false and fabricated data. This is not an isolated instance of false or fabricated data, amounting to an honest mistake. Rather, this conduct occurred on an ongoing basis, and pervaded numerous publications and grant applications. Respondent simply has not disputed the overwhelming evidence that he repeatedly submitted false images and data; rather, he only challenged who created the images and data. The data flaws in Respondent’s grant applications and journal articles were significant and should have been apparent, had Respondent taken care to review his data prior to reporting the data in his submissions. Even if, assuming for summary judgment, Respondent did not create the falsified and fabricated images, Respondent was reckless in repeatedly adopting those flawed images as his own and reporting them in grant applications and journal articles. (Sauer 2017, 16)

Thus, the Sauer opinion supports the concept that recklessness in research misconduct can be equated to “indifference to the truth” or “without regard for the truth,” and that factors weighing in support of a recklessness finding include the ongoing nature, pervasiveness, and obviousness of the misrepresentations.

ORI v. Kreipke

ORI v. Kreipke involved findings of research misconduct against Dr. Kreipke, the same individual discussed above. ORI found that Dr. Kreipke “recklessly caused or permitted twenty-three (23) instances of research misconduct in his three (3) grant applications, two (2) articles on which he was the first listed author, and two (2) posters on which he was the first listed author” (Office of the Secretary, HHS 2018a, 38316). Dr. Kreipke appealed these findings.

In a 2018 decision, the ALJ denied Dr. Kreipke’s appeal. The ALJ specifically found that Dr. Kreipke acted recklessly, rather than intentionally or knowingly, and discussed extensively how the evidence supported a recklessness finding (Kreipke 2018). Relying upon Merriam-Webster Dictionary and Black’s Law Dictionary, the ALJ defined the intent standards as follows: “[I]ntentionally means one acts with the aim of carrying out the act. Knowingly means that one acts with knowledge and information and awareness of the act. Recklessly means one acts without proper caution despite a known risk for harm” (14). The ALJ further commented that for recklessness, ORI needed to demonstrate, by a preponderance of the evidence, that the respondent “[u]sed materials without exercising proper care or caution and disregarding or showing indifference to the risk that the materials were false, fabricated or plagiarized thereby causing harm to the integrity of the research process or waste of public funds, the basis for regulating research misconduct stated in 42 C.F.R. § 93.100” (14).

The Kreipke standard is similar to that suggested in Sauer, though it specifies that the indifference should be to the risk of falsification, fabrication, or plagiarism of the specific research records in question in addition to the broad “repeated, pervasive indifference to the truth” concept suggested in Sauer. Moreover, Kreipke held that in order to find research misconduct, a respondent must also fail to exercise proper care or caution with regard to the research.

Based on these standards, the ALJ in Kreipke introduced two distinct tests, one objective and another subjective, either of which can be applied to determine recklessness (Kreipke 2018, 83).

• Objective: “Therefore, including data, images, and other materials without validation of the accuracy of the information constitutes failure to exercise proper care or caution and disregard or indifference to the risk for potential false information, and a violation of the responsibility institutions and researchers accept to protect PHS funds and the integrity of PHS research.”

• Subjective: “Based on his knowledge of the state of his laboratory and personnel situations, it was reckless for Respondent to simply assume that materials placed in his grants, articles, and posters were reliable.”

Applying the recklessness standard to the facts of the case, the ALJ found that ORI had established a sufficient prima facie case that Dr. Kreipke “was aware of disorganization and lack of record keeping when he joined the laboratory and for which he became leader … [and] was aware that virtually everyone that worked in the laboratory had access to all the computers and laboratory notebooks.” Citing these facts, the ALJ concluded that “it was reckless for [Dr. Kreipke] to simply assume that materials placed in his grants, articles, and posters were reliable.”

In sum, as a legal standard, Kreipke holds that recklessness must be considered in the context of the researcher and their individual knowledge and that, further, the researcher must both fail to exercise proper care or caution and disregard or be indifferent to the risk of falsification, fabrication, or plagiarism in order to reach a recklessness finding. In terms of practical takeaways, the Kreipke opinion helps illustrate the importance of maintaining defined systems of data management, review, and preservation – basic elements of scientific practice that, if neglected, would tend to tempt laboratory members who understand that the risk of their misconduct being identified is lessened when original research records do not exist, thus increasing “the risk for potential false information” being inserted into a published paper or presentation.

In the immediate aftermath of Kreipke, an HHS spokesperson, on behalf of ORI, endorsed the Kreipke standard. In an article published in Report on Research Compliance in September 2018, the spokesperson stated that Kreipke was “thorough and careful,” that “ORI … believes the research community is likely to find it useful and educational,” and that the ruling “provides … a judge’s thoughts on how to define the terms … [and] will be helpful to ORI in deciding how to approach future cases” (Health Care Compliance Association 2018). ORI also endorsed this case in its Fall 2018 newsletter (Office of Research Integrity 2018). Recently, ORI has pointed to Kreipke when responding to informal requests by institutions as to how they should apply the recklessness standard in pending cases.³

ORI v. Srivastava

In ORI v. Srivastava, ORI found that Rakesh Srivastava, Ph.D., “intentionally committed research misconduct by submitting to NIH a grant application that included plagiarized words” (Office of the Secretary, HHS 2018b, 57739). Dr. Srivastava appealed these findings. In a 2018 decision, the ALJ denied Dr. Srivastava’s appeal. The ALJ opinion is consistent with and builds upon the opinion issued in Kreipke, providing that to establish recklessness, ORI needed to show that the respondent “used materials without exercising proper care or caution and disregarding or showing indifference to the risk that the materials were false, fabricated or plagiarized” (Srivastava 2018, 12).

Recklessness under other sets of federal research misconduct regulations

Federal funding agencies are bound by the 2000 mandate from OSTP to conform their research misconduct policies and procedures to the Federal Research Misconduct Policy, including the instruction that “[t]he misconduct be committed intentionally, or knowingly, or recklessly.” In this section, we consider how recklessness is used in the research misconduct context by the National Science Foundation (“NSF”) and National Aeronautics and Space Administration (“NASA”), given that these two agencies have discussed recklessness using slightly different articulations that help supplement and provide additional color to the framework set forth under the Federal Research Misconduct Policy and Part 93.

National Science Foundation

NSF regulations at 45 C.F.R. § 689.2(c) provide that a finding of research misconduct requires: “(1) There be a significant departure from accepted practices of the relevant research community; and (2) The research misconduct be committed intentionally, or knowingly, or recklessly; and (3) The allegation be proven by a preponderance of evidence.” Like ORI, NSF’s regulations do not define intentional, knowing, or reckless. However, unlike ORI, the NSF Office of Inspector General (“NSF OIG”) has prepared guidance on the applicable intent standards, titled “Assessing Intent in Research Misconduct Investigations.” In this guidance, NSF OIG explains that “[t]he assessment of intent by institution investigation committees can be a challenge, so we developed this document to facilitate such assessments” (NSF Office of Inspector General n.d., 1). NSF OIG then sets forth the following definition for reckless:

Reckless: The subject used materials with a lack of proper caution and/or showed indifference to the risk that the materials may be false, fabricated, or plagiarized. The subject took a risk with materials without thinking or caring about the consequences of the action, even if the subject did not fully realize the risk.

NSF OIG notes that these definitions are influenced by Black’s Law Dictionary, the Merriam-Webster online dictionary, and the Srivastava ALJ decision. We note that in Srivastava, the ALJ likewise cited Black’s Law Dictionary and Merriam-Webster. Moreover, the ALJ in Kreipke cited both dictionary definitions and applied a similar definition for recklessness.

National Aeronautics and Space Administration

NASA regulations at 14 C.F.R. § 1275.105 state that a research misconduct investigation report should “include recommended findings as to whether research misconduct has occurred,” noting that such a finding requires that:

1. there be a significant departure from accepted practices of the relevant research community for maintaining the integrity of the research record;

2. the research misconduct be committed intentionally, knowingly, or in reckless disregard of accepted practices; and

3. the allegation be proven by a preponderance of evidence.

The NASA regulations thus adopt a standard comparable to “recklessness,” but identified as the more specific “reckless disregard of accepted practices.” In theory, the NASA standard is narrower than the “misconduct committed recklessly” standard under the Federal Research Misconduct Policy and the verbatim standard utilized under Part 93, as it seemingly would compel a decision-maker to identify an accepted practice and then determine, by a preponderance of the evidence, that the falsification, fabrication, or plagiarism constitutes reckless disregard of that particular practice. However, it is difficult to identify a scenario in which this standard would result in a different outcome than the standard detailed under the Federal Research Misconduct Policy and Part 93, as those texts also require decision-makers to determine whether there has been “a significant departure from accepted practices of the relevant research community.”

Recklessness in sampling of other legal contexts

A “recklessness” standard is used in legal contexts other than research. To illustrate the use of “recklessness” elsewhere in law, we summarize briefly the application of a recklessness standard under some of these additional contexts: the federal False Claims Act (“FCA”), Disqualification of a Clinical Investigator by the Food and Drug Administration (“FDA”), and federal securities law regulated by the U.S. Securities and Exchange Commission (“SEC”). In short, while the concept of recklessness is widely used in connection with each of these legal doctrines and the standards used in these contexts are useful for thinking about how best to craft a definition for recklessness, none of these recklessness doctrines provide guidance that is more detailed or helpful than Kreipke for purposes of establishing a “recklessness” standard in research misconduct proceedings.

False claims act

The FCA, 31 U.S.C. § 3729, applies to any person who “knowingly,” among other things, makes false or fraudulent claims. The FCA includes reckless in the definition of “knowing,” stating:

The terms “knowing” and “knowingly” … (A) mean that a person, with respect to information … (i) has actual knowledge of the information; (ii) acts in deliberate ignorance of the truth or falsity of the information; or (iii) acts in reckless disregard of the truth or falsity of the information; and (B) require no proof of specific intent to defraud.

The FCA provides that recklessness can either be evidenced by deliberate ignorance or reckless disregard of the truth. We note that the “reckless disregard” prong of the definition is very similar to the definition of recklessness used in the various ALJ opinions described above.

The definition of “knowingly” under the FCA is of relevance to research institutions that are reviewing allegations of research misconduct relating to federally funded work for a more specific reason as well: in certain circumstances, allegations of research misconduct could also give rise to liability under the FCA itself. If a research misconduct investigation reveals instances of falsified or fabricated data that were used or relied upon in a federal grant application or progress report submitted by the institution, the institution needs to assess whether the use of or reliance upon those data in a submission to the funding agency could give rise to liability under the FCA. The potential liabilities under the FCA arising from research misconduct can be significant. Perhaps the most prominent example is that in 2019, Duke University agreed to pay $112.5 million to settle FCA allegations related to research misconduct in thirty NIH and Environmental Protection Agency grants (Department of Justice 2019).

FDA: Disqualification of a clinical investigator

In 2012, FDA amended its regulations at 21 C.F.R. § 312.70 to expand the basis on which a clinical investigator can be disqualified from conducting any clinical investigation that supports “an application for a research or marketing permit for products regulated by FDA.” Like the PHS research misconduct regulations, falsity of submitted information is at issue under these regulations. FDA considers the submission of false information to be grounds for disqualification if the conduct is either repeated or deliberate. In the 2012 Federal Register commentary, FDA explains that “reckless disregard” can be considered to be “deliberate” and thus can lead to disqualification of a clinical investigator:

The term “deliberately” includes conduct that is “willful” as well as conduct demonstrating reckless disregard … Likewise, an investigator who shows a reckless disregard for whether their conduct may result in a regulatory violation may be found to have deliberately violated the regulations … Therefore, to sustain a finding of repeated or deliberate submission of false information, FDA must show that the clinical investigator repeatedly submitted to the sponsor or to FDA false information, whether in a single study or in multiple studies, or submitted false information to the sponsor or FDA knowingly or willfully or with reckless disregard for the truthfulness of the data submitted. (Food and Drug Administration, HHS 2012, 25354–55)

Notably, the language used by FDA – “reckless disregard for the truthfulness of the data” – is similar to the various iterations of the definition described in earlier sections.

Securities law

The SEC promulgated Rule 10b-5 under Section 10(b) of the Securities Exchange Act of 1934, which authorizes the SEC to regulate securities fraud. The regulation at 17 C.F.R. § 240.10b-5 states:

[I]t shall be unlawful for any person … (a) [t]o employ any device, scheme, or artifice to defraud, (b) [t]o make any untrue statement of a material fact or to omit to state a material fact … or (c) to engage in any act, practice, or course of business which operates or would operate as a fraud or deceit upon any person, in connection with the purchase or sale of any security.

The Supreme Court has established that to find liability under Rule 10b-5, one must find evidence of scienter, defined as “a mental state embracing intent to deceive, manipulate, or defraud” (Ernst and Ernst 1976, 193–194, n.12). Lower courts have recognized that the scienter standard may be satisfied based on a recklessness standard, and multiple courts have adopted a definition of recklessness set forth by the Seventh Circuit:

Reckless conduct may be defined as … highly unreasonable [conduct], involving not merely simple, or even inexcusable negligence, but an extreme departure from the standards of ordinary care, and which presents a danger of misleading buyers or sellers that is either known to the defendant or is so obvious that the actor must have been aware of it. (Sundstrand Corporation v Sun Chemical Corporation 1977, 1045)

While the SEC may define recklessness differently than we have seen in other contexts, the SEC standard nevertheless deploys a two-prong test, with both objective and subjective parts, similar to the Kreipke decision described above.

Proposed approach for evaluating recklessness

A consistent focus throughout the recklessness precedent summarized above is an emphasis on indifference to or disregard for the truth of the matter being asserted (except for the NASA regulations, which focus on recklessness in regard to a specific research-related practice). Consistent with this prevailing practice, one possible standard for recklessness derived from Kreipke that institutions might find functional and legally reasonable could be stated as: In order to find research misconduct based on recklessness, a respondent must have disregarded or been indifferent to the risk that the materials were false, fabricated, or plagiarized.

However, while this proposed formulation is more specific than the mere reference to “recklessly” under 42 C.F.R. § 93.104(b), it is still not sufficiently specific to ensure consistency in applying “recklessness” across different research misconduct cases, each with its own distinct facts and circumstances. Building on the more detailed guidance offered in Kreipke and other cases, institutions might find it useful to ask standard, more detailed questions when assessing a recklessness issue during a research misconduct proceeding. The questions below represent a suggested two-part test, under which institutions must answer both questions in the affirmative (based on the preponderance-of-evidence standard set forth under Part 93) in order to reach a finding of research misconduct:

Part 1: Did the respondent include false, fabricated, or plagiarized data without verifying the accuracy of the information presented?

Part 2: Did the respondent take appropriate and sufficient action to ensure the integrity of the data presented and to mitigate the risk that data were false, fabricated, or plagiarized?

These questions are consistent with the standard articulated in Kreipke. Specifically, Part 1 of the test reflects the objective standard set forth in Kreipke (i.e., were data included that were not verified?), while Part 2 of the test reflects the subjective standard set forth in Kreipke (i.e., given the lab environment and the nature of the research, were precautionary measures taken?). However, we think that the above-defined test offers a more concise articulation of “disregard or indifference to the risk” in the context of the respondent’s particular research environment.

We note that while Part 1 of the above test often results in a straightforward analysis, Part 2 can be more difficult to assess, requiring the relevant decision-maker to evaluate the unique circumstances of the respondent’s particular research environment and of the work in question and compare those circumstances to generally accepted practices in the field. As such, the factors set forth in Table 1 might be used to evaluate whether a respondent was reckless instead of simply negligent.

Table 1. Factors to consider in recklessness determination.

	Factor
1.	Widespread perception in the field that the specific category of research or the specific laboratory culture itself suggests enhanced risk for falsification, fabrication, or plagiarism.
2.	Repeated or sustained nature of permitting, tolerating, and/or ignoring of cultural conditions or lab environments that would tend to lead to problems in the conduct and reporting of research.
3.	Lack of controls that otherwise are typical of laboratories and research efforts in the field, including, but not limited to:
	(i) Failure to utilize available tools to confirm the veracity of available data, or otherwise implement data integrity systems.
	(ii) Failure to implement systems to manage, maintain, and store research records and data.
	(iii) Failure to train supervisees appropriately on data integrity practices.
4.	Lack of attentiveness to or use of well-recognized oversight mechanisms and practices (e.g., review of raw data in laboratory meetings, validation of data by multiple laboratory members before submission for publication).
5.	Lack of remedial action in response to previous allegations or concerns relating to falsification, fabrication, or plagiarism.
6.	Context of the laboratory and its staff (e.g., structure of laboratory supervision, frequency of laboratory meetings, oversight practices, policies and procedures on data collection, management, sharing, and retention).

During the investigation phase of a research misconduct proceeding focused on the culpability of a laboratory director or other research supervisor, the factors outlined in Table 1 can be assessed through a primary focus on three key factors: (1) a respondent’s laboratory management practices, including data collection, management, sharing, and retention practices and, more generally; (2) the nature and frequency of the respondent’s oversight of laboratory members’ daily scientific work; and (3) the ready availability of clearly, accurately labeled and contemporaneously preserved source data and whether those data support the integrity of the published or presented findings. In cases in which it appears that the level and rigor of supervision by respondents will be significant to a committee’s findings regarding recklessness, a review of available research records and discussion with the respondent regarding their supervisory practices may only provide limited practical insight. In many cases, it may be necessary to seek interviews of current and former laboratory members in order to assess more fully and accurately the environment of the laboratory in which the falsification, fabrication, or plagiarism allegedly occurred.

If a finding in support of recklessness has been made utilizing the two-part test described above, then, per 42 C.F.R. § 93.104, the fact-finder must also determine whether the behavior in question is “a significant departure from accepted practices of the relevant research community” in order to reach a finding that the respondent committed research misconduct. In our experience, if a research misconduct investigation report is thorough and explains its analysis of the respondent’s research and lab management practices in detail (e.g., why the decision-maker found or declined to find that the respondent was reckless), then the question of “significant departure from accepted practices” typically becomes redundant, as the answer lies in the analysis of the respondent’s practices contained in the report and supports a finding that the respondent’s reckless behavior also constitutes a significant departure from accepted practices. However, asking this question is nevertheless an important final procedural safeguard for respondents, encouraging investigation committees not only to think about the recklessness standard based on the facts of the individual case but also to ask whether similarly situated researchers might have reasonably acted in a similar fashion with respect to their oversight practices. It is conceivable that, in rare circumstances, an institution could conclude that a respondent was “reckless” but that the actions did not constitute a significant departure from accepted practices and, therefore, that the respondent did not commit research misconduct.

Another more radical approach here would be for research misconduct regulations (and institutional policies implementing those regulations) to be amended to separate research misconduct based on intentional or knowing falsification or fabrication of data, or plagiarism, from a “lesser charge” of grossly negligent failure to supervise, or grossly negligent failure to verify, research. In our experience, even when there may be very direct evidence of serious and persistent supervisory failure – as, for example, in research settings in which there are no standards or directions for maintaining original data – there is a reticence of investigation committees to find the scientific supervisor to have acted “recklessly;” these committees are reticent to impose a judgment of “research misconduct” on a supervising researcher for their nonfeasance. Establishing a separate charge, also cognizable in the existing research misconduct system, of “failure to supervise or validate” according to a standard of “gross negligence” could allow committees to find a supervising researcher to have committed this sort of serious professionalism violation, without grouping that researcher with those who have intentionally or knowingly falsified or fabricated data. This change would require regulatory reform, and not mere guidance, but it is offered as a potential strategy to facilitate inquiries into supervisory behaviors that could have been a directly and significantly causative factor in data falsification, fabrication, or plagiarism.⁴ The peril of this entirely alternate approach is that without an extremely detailed, carefully constructed, narrow definition, it could result in a multiplicity of charges and allegations just based on poor research supervision and could overwhelm an already highly stressed research misconduct infrastructure.

Discussion: Implementation of recklessness standard and implications for research integrity training and education

This article aims to assist institutions that conduct research misconduct proceedings, including inquiry and investigation committee members and deciding officials, by providing more detailed guidance on the recklessness standard than has previously been offered in this context, and by offering a proposed framework for assessing recklessness in a misconduct case. We also hope that these proposals will facilitate additional discussion among stakeholders in the research community about the recklessness standard and lead to the development of additional guidance and clarity. ORI would provide a service to the research community by issuing guidance and case studies detailing how research institutions and ORI itself have appropriately applied the recklessness standard to various research misconduct cases, particularly those involving the work of a respondent who supervised the work in question. Professional societies could also develop guidance on best practices for laboratory oversight and management, and strategies for fostering a laboratory culture conducive to research integrity, particularly when generally accepted practices vary by discipline or field and method of study; such guidance would tend to define what is and is not “reckless” conduct in the area of laboratory supervision.

Additional guidance regarding “recklessness” in the context of research oversight and laboratory supervision would certainly improve the research misconduct investigation process, but we hope that it would also have a cascading effect and ultimately lead to public policy changes. The first policy aim is greater fairness for researchers. Clear and specific guidelines are extremely important to minimize arbitrary and inconsistent outcomes. To be specific, even researchers with the most diligent practices and strongest commitment to ethical standards are sometimes caught up in research misconduct proceedings involving instances of falsification or fabrication by their colleagues. In such cases, the diligent researcher may rightly think that the possibility of a recklessness finding is not remote, given that an institution is compelled to reach a finding of research misconduct if it concludes that it is “more likely than not,” i.e., greater than 50% likelihood, that the respondent’s failure to identify and prevent an instance of falsification or fabrication constituted “recklessness.” The preponderance of evidence and recklessness standards serve their own public policy aims, helping to ensure accountability by researchers even if there are gaps in information that render it impossible to understand exactly what went wrong in the conduct and reporting of scientific experiments.

A related policy aim is that clearer and more predictable standards will allow institutions to provide more effective training and guidance to their researchers regarding appropriate standards for the supervision of research. More effective training and guidance will, among other aims, help mitigate risk of falsification and fabrication occurring under a researcher’s watch. Many research institutions are highly decentralized organizations that support many diverse types of research. As such, it may be impossible or highly impractical to implement universal standards for recordkeeping and data storage, laboratory meetings, review of primary data, and other essential functions of research oversight. Nevertheless, certain universal elements can be emphasized. Researchers who supervise others must have a system for: (1) reviewing primary data generated by others who are contributing to a project; (2) ensuring that original data are retained for time periods required by grant terms, laws, and institutional policies; and (3) storing and retrieving research records and materials. These systems and the supervisor’s expectations regarding each laboratory member’s responsibilities must be known to all members of the laboratory. Enhanced guidance and discussion on the recklessness standard may feed into routine, structured educational efforts being carried out by institutions (e.g., RCR training). For example, guidance on recklessness can be embedded in case studies within the relevant training, to emphasize the importance of maintaining consistent, reliable and transparent data integrity protections. These trainings fit within the key topics that NIH requires for RCR training, including “mentor/mentee responsibilities and relationships,” “data acquisition and analysis,” “laboratory tools (e.g., tools for analyzing data and creating or working with digital images),” “recordkeeping practices,” “research misconduct and policies for handling misconduct,” “responsible authorship and publication,” and “the scientist as a responsible member of society” (National Institutes of Health 2022).

In summary, there must be clear expectations that the responsibility for data integrity includes not only the work of those conducting experiments, but also those who ultimately supervise those experiments, and without whose reputation the work may not have been funded and published. In turn, the expectations of what is reasonable scientific supervision, and when that supervision is so lacking that it becomes “reckless,” need to be made clear, so that supervising scientists have fair notice of their duties and can conform their own conduct accordingly. In this way, guidance on how to interpret and apply the “recklessness” standard in research misconduct proceedings, if clear and reasonable, can help to re-order and strengthen the supervision of research.

Disclosure statement

Mark Barnes, J.D. LL.M., is a partner, and Minal Caron, J.D., and Sarah Dohan, J.D., are associates, at Ropes and Gray LLP, an international law firm that represents multiple entities in the health care and academic sectors, including by advising academic medical centers and institutions of higher education on research misconduct matters.

Following acceptance of this manuscript and prior to publication, ORI published three additional case summaries, including one related to recklessness. In this recent 2023 case summary, ORI found that Andrew Dannenberg, M.D., 'recklessly reported falsified and/or fabricated Western blot image data that were reused, with or without manipulation to conceal their similarities, and falsely relabeled as data representing different experiments or proteins in sixty (60) figure panels included in twelve (12) published papers' (Office of the Secretary, HHS 2023, 62804). As in the two case summaries described above, this case summary also does not explain how a finding of recklessness was reached. However, the case summary suggests that the failure to maintain reliable data was a factor, noting, 'In the absence of reliable image and numerical data, the figures, statistical analyses, and related text also are false' (62804).

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.

Notes

1. We refer to “laboratory supervision” at many points throughout this article for simplicity and because “laboratory” is often used to refer to a research group even if the research conducted by the group does not occur in a traditional laboratory setting. For example, even NIH’s own guidance on RCR training focuses on concepts like “laboratory tools (e.g., tools for analyzing data)” even though such tools, and the RCR training itself, extend to many diverse research environments other than a laboratory setting (e.g., clinical research) (National Institutes of Health 2022). All types of research – not just research being carried out in a laboratory setting – are susceptible to lapses in oversight by the senior members of the research team, and these lapses can give rise to data integrity concerns and, potentially, findings of research misconduct based on poor supervision. As such, notwithstanding our use of the term “laboratory” throughout this article when referring to the supervision of a research group, the discussion of recklessness in this article is not limited to types of work performed in a laboratory environment, and includes all disciplines in which research can include included falsified or fabricated data, or plagiarized work.

2. All ORI case summaries are published in the Federal Register. However, ORI sanctions are issued for a time-limited period, and ORI typically removes case summaries from its website once the sanction period passes. The authors did not conduct an exhaustive review of ORI case studies published in the Federal Register for purposes of preparing this article; we reviewed only those summaries available on the ORI website. However, we have reviewed the ORI case studies over a period of many years and have always reached the same conclusion: specifically, that the case studies do not provide detailed information on ORI’s interpretation of “recklessly” under Part 93.

3. This statement is derived from the personal knowledge of the authors, based on their correspondence with ORI in specific cases that are subject to ORI oversight.

4. Alternatively, research institutions could adopt into their institutional professional standards explicit expectations regarding research supervision that, if breached, would give rise to the possibility of disciplinary action pursuant to robust internal faculty or staff disciplinary proceedings, not processed under research misconduct at all.