Criminalizing medical research fraud: Towards an appropriate legal framework and policy response

Extent and nature of the problem

A 2009 systematic review on international data suggested that 2% of researchers would admit to having fabricated or falsified research data. ¹⁹ Journal paper retraction rates have increased by 10-fold in recent years, with 43% of retractions being related to fraud or suspected fraud. ²⁰ Though arguably rare by incidence, MRF is far from a negligible problem as there are now over seven million researchers in the world, reading and publishing in over 25,000 scientific journals yearly. ²¹ In the UK, one in seven researchers admitted knowledge of colleagues ‘inappropriately adjusting, excluding, altering, or fabricating data’. ²² Ten percent of newly appointed consultants from seven National Health Service (NHS) hospital trusts described similar experiences; 6% also admitted to past personal research misconduct. ²³ MRF is not a foreign anomaly. ²⁴

The above figures should be interpreted cautiously. Paper retraction rate, for one, is an unreliable surrogate marker of research behaviour as it only captures discovered acts of serious misconduct; ²⁵ plenty may have gone undetected or unreported. ²⁶ On the other hand, the apparent rise in the incidence of MRF could well be due to improved knowledge, awareness and detection. ²⁷ A comprehensive assessment of the true extent of MRF is therefore difficult to come by. What is particularly disconcerting though is when supposedly dependable official figures have also turned out to be misleading. ²⁸ In 2017, data obtained by reporters under the Freedom of Information legislation ²⁹ revealed over 300 MRF-related allegations at UK Russell Group universities between 2011 and 2016. This contrasted sharply with the much lower figure of 33 cases reported by the Research Council UK (RCUK). ³⁰ The conspicuous lack of transparency and accountability is worrisome. It emphasizes the importance of an official body vested with the authority to review cases and sanction institutions that do not comply with agreed rules. ³¹

The complexity of MRF is well recognized. Early UK cases tended to involve individual researchers publishing faked articles on their own but the scientific landscape has since changed dramatically. ³² Laboratories and research groups are now much larger and multi-institutional in nature. Increased commercialization may give rise to conflicts of interest and influence research practice ³³ whereby commercial companies may allegedly suppress unfavourable findings or employ ‘ghost-writers’ to publish manipulated data. ³⁴ The problem could be particularly prominent in the US, where a quarter of investigators have industry affiliations, ³⁵ and where significant associations between pro-industry research conclusion and industry sponsorship have been demonstrated. ³⁶ The situation is probably less acute (and therefore receiving less attention) in the UK, where less than 5% of research funding received by universities comes from the private sector. ³⁷

The Internet offers new opportunities for improper research activities. Predatory journals that profit from publication fees but do not subject submissions to peer review could be enabling the publication of fraudulent research works. ³⁸ Many operate an open-access model and deceive readers and researchers with false claims of review practice; ³⁹ others may hijack websites of reputed journals and ascribe undue credit to fraudsters and their works. ⁴⁰ There is, in addition, an emerging practice of peer-review fraud where authors would recommend ‘expert reviewers’ to journal editors and provide email addresses that route messages from the journal back to the researchers themselves. ⁴¹ Left undeterred, these activities can quickly corrupt the literature and erode trust.

From a global perspective, the rise in international research collaborations means that a single piece of work may now involve parties from multiple jurisdictions. MRF committed under these situations may conceivably trigger offences that were not originally designed for such use, and have probably never been used, in dealing with research misconduct. Liabilities and harm may surface under different jurisdictions at different times, and prosecutors may find themselves in uncharted territories, some of which may well be foreign waters. Even an experienced agency such as the Office of Research Integrity in the US has found it difficult to pursue allegations across borders. ⁴² MRF has thus evolved from an obscured academic issue that concerned a minority of scholars and scientists to one that has bearing for an entire country if not the world at large.

Consequences of MRF

MRF is distinct from academic fraud in other disciplines for its potential impact on health and healthcare, and an increased readiness to translate novel findings into clinical practice means that ensuing harm can be disseminated rapidly. ⁴³ The extent to which MRF causes actual medical injury is unknown and is probably unknowable. Lying deep beyond the ‘tip of the iceberg’ could be examples of MRF that have already caused, or have the potential to cause, incalculable damages that will never be discovered. At the same time, the public’s growing sense of involvement in high-profile cases such as that of Anjan Kumar Banerjee, ⁴⁴ Thorsten Hagemann, ⁴⁵ Robert Ryan ⁴⁶ and Andrew Wakefield ⁴⁷ has put MRF firmly on the public agenda.

And a global agenda it is, too. For example, Wakefield’s false claim about the link between measles, mumps and rubella (MMR) vaccine and autism was a powerful fuel for the ‘anti-vaccination movement’ that led to a significant and persistent decline in vaccinations rates worldwide. Multiple outbreaks causing deaths continue to occur in the US, UK and other European countries almost 20 years after his paper’s publication. ⁴⁸ The fact that many fraudulent works are published in high-ranking journals also means that they are prone to be used for formulating clinical practice guidelines with potentially serious and long-lasting consequences. ⁴⁹ A case in point is the discredited studies by Don Poldermans on the perioperative use of beta blockers that had allegedly caused over 10,000 deaths in the UK through its influence on the European Society of Cardiology’s guidelines. ⁵⁰ Even basic science research or research studies related to psychology, healthcare policy, economics, climate change, that generally do not have immediate clinical impact can still cause long-term damage. ⁵¹ A collective responsibility exists to deter and prevent MRF.

The world spends over US$1000 billion yearly on funding research, the allocation of which is supposedly based on merits and trust. ⁵² MRF undermines this trust and misdirects the use of public money. The investigation of a suspected case of research fraud itself can be costly, estimated at around US$500,000, ⁵³ while honest investigators who unknowingly base their works on faked researches stand to lose valuable time and resources. Publishers, journals and their readers also suffer, albeit in less tangible ways. Reputational losses to academic institutions and science are no less important. Volunteers who participate in clinical studies put themselves at risk, and MRF violates their rights and denies the value of their altruistic acts. ⁵⁴ The number and varieties of stakeholders are indeed staggering, rendering MRF a far more complex problem to manage than many other fraudulent or criminal activities.

It is apt to mention here that MRF may also have implications for the administration of justice itself. In medical negligence litigation, the determination of standards of care and causation relies heavily on expert witness opinions. ⁵⁵ The latter commonly refer to clinical practice guidelines, which are in turn based on information generated from prior researches. ⁵⁶ Fraudulent and misleading research findings could therefore indirectly compromise judicial proceedings. Conversely, an exposure of research fraud could, conceptually speaking, invalidate expert opinions and judicial decisions.

In the aforementioned case of Don Poldermans, subsequent reviews found that clinical recommendations would have been completely different once his data were excluded. ⁵⁷ The author is unaware of any allegations brought in that connection, but one can readily entertain the possibility that some doctors could have been found liable in negligence, scientifically wrongly, because of Poldermans’ works. The ensuing arguments might turn upon whether standards of care should have been determined, through expert opinion, by the best available information at the time even when it was based on fraudulent research, or the correct information that should have been available had there been no faked data. In Wakefield’s case, a group action involving hundreds of MMR vaccine litigants had to be dropped after Wakefield’s study was discredited. It cost an estimated £15 million in legal aid and one of the litigants later sued his law firm for pursuing a ‘hopeless claim’. ⁵⁸ This aspect of MRF has not attracted much attention in the literature and deserves further scholarly investigation.

Current approaches towards MRF

In the UK, the Department of Business, Energy and Industrial Strategy is responsible for policy decisions on research. Its activities, where there are implications for science, engineering, technology and research, are scrutinized by the POST. The latter’s 2011 report on ‘Peer review in scientific publications’ led to the establishment of the Research Integrity Concordat by Universities UK. ⁵⁹ Apart from setting out standards of research practice and performing audits, ‘the Concordat’ makes compliance a condition for funding by public bodies such as the Higher Education Funding Councils and the Wellcome Trust. ⁶⁰ Under the Care Act 2014, the Health Research Authority was established to provide statutory oversight within the NHS. ⁶¹

Other independent bodies such as the UK Research Integrity Office (UKRIO) and the COPE as well as professional organizations and academic institutions are there to provide guidance, training and supervision. Journal editors serve as gatekeepers by reviewing manuscripts, while fellow researchers may conduct post-publication scrutiny. The GMC adjudicates and sanctions serious acts of research misconduct by registered medical practitioners. ⁶² At the supranational level, the European Code of Conduct for Research Integrity laid down relevant principles and guidance, but the code is not legally binding. ⁶³

Criticisms of the above set-up are well rehearsed. ⁶⁴ Of particular concern is the absence of an official body in the UK, available in the US ⁶⁵ and Denmark, ⁶⁶ that is dedicated to the management of research misconduct (and which might also explain why legal proceedings have been less common in the UK than in the US). Indeed, organizations such as the Concordat ⁶⁷ and the British Medical Association ⁶⁸ have continued to express reservations towards further direct control by the government or additional professional regulatory processes, citing the importance for stakeholders to enjoy ‘flexibility’ and be ‘innovative’ in fostering research integrity. But the very idea that science can somehow self-police and correct itself is subject to challenge. ⁶⁹ Time constraints, cost and the lack of systematic training means that peer-review processes may detect plagiarism effectively but not data falsification or fabrication. ⁷⁰ Potential whistle-blowers may be deterred for fear of career repercussions or legal liabilities, ⁷¹ while journal editors may fail to respond to complaints. ⁷² The investigations of suspected cases can be further hindered by institutional culture and concerns about reputational damage. ⁷³ According to a 2015 survey, for example, only a minority of universities that subscribed to the UKRIO had published summaries of their investigations into research misconduct; a much lower figure was found with institutions not party to it. ⁷⁴

The effectiveness of non-regulatory measures is similarly questionable. ⁷⁵ Although education and training in research integrity has been shown to influence ethical knowledge and awareness, there are no data to show that it actually reduces research misconduct; ⁷⁶ it could in fact be counterproductive as students of courses on responsible research conduct may become overconfident in their abilities in handling ethical problems. ⁷⁷ Even if training can diminish honest errors, sloppy practices and deliberate acts of misconduct to a certain extent, it is unlikely to deter those who are determined to deceive for personal gain against all odds; nor would it necessarily stop companies, driven by commercial interests, from engaging in data manipulation. ⁷⁸ The overall picture is one of inadequate prevention, deterrence and monitoring.

Such challenges are not unique to the UK as other countries continue to struggle with an effective and politically acceptable model for detecting or scrutinizing MRF. ⁷⁹ Of note is the discontent among researchers against legislative control that is often linked with research funding allocation in the US, where political and democratic arguments held that aggressive government interventions in matters of research were not only appropriate but also necessary. ⁸⁰ The allegation that such an interventional approach may be misused to ensure conformity of research findings with bureaucratic preferences further intensified the debate. ⁸¹ Notwithstanding, the decision by the UK government to launch a formal inquiry into the matter is strongly indicative of a perceived need for change. ⁸² Meanwhile, journal editors, ⁸³ researchers ⁸⁴ and legal scholars ⁸⁵ have called for the introduction of criminal sanctions with a view to deter and punish fraudulent research activities. ⁸⁶

Justifying criminalization

The justification for criminal law intervention in the present context is a complex issue, involving, inter alia, fundamental questions regarding the definition of crime, its constituent elements and the legitimacy for the criminal law to reach into what is perceived by some as the ‘private sphere’ of science. ⁸⁷ An in-depth debate is outwith the scope of this article. It suffices to say that some authors considered MRF to be ‘characteristic of other conducts that are treated criminally, and that the scientific context should not insulate them from scrutiny’. ⁸⁸ Currently, the most severe non-legal sanctions are those imposed by the GMC, and these have been criticized for being relatively light and disproportionate to the harm or potential harm caused. ⁸⁹ Importantly, the GMC only has jurisdictions over registered medical practitioners but not other clinical investigators or basic scientists. Although the Science Council operates a registry for the latter, registration is voluntary and the agency lacks sanctioning power. ⁹⁰

Others in support of criminalization have emphasized the symbolic role of the criminal law in restoring trust and the fact that involvement of the criminal justice system can ensure due process and cost-effective investigations. ⁹¹ The police, for example, are better resourced and more experienced than journals or academic institutions in gathering evidence, while the Specialist Fraud Division under the CPS is dedicated to the pursuing of serious, complex and difficult cases.

Opponents have countered that criminalization would undermine trust, have a ‘chilling effect’ on science, and that it simply ‘will not work’. ⁹² Detection is probably the weakest link in the current management of MRF, but it is also where criminalization has the least to offer in that a fraudulent act cannot be pursued without being detected and reported in the first place, and the prospect of having to give evidence in an open court could deter whistle-blowers from coming forward despite the availability of a certain degree of legal protection. ⁹³ Furthermore, although incarceration would stop a researcher from publishing for a while, it might not prevent ensuing medical harm, since journal publications can influence medical practices for many years after their retraction. ⁹⁴ ‘Minor’ cases of MRF research fraud, according to some authors, are best dealt with by academic institutions or the GMC. ⁹⁵ The critical question is how to determine ‘seriousness’ in something as complex as MRF. This will be discussed later.

It must be emphasized that criminalization is but one of many ways to improve the current situation. Criminal sanction is not supposed to, and should not be expected to, replace other measures; all measures should aim at complementing each other. But the fact remains that MRF is largely untouched by the criminal law in the UK. The POST has acknowledged the controversy but given no indication of the direction of travel. ⁹⁶ To further the present inquiry, it is helpful to review some of the relevant legal instruments in the UK and the developments in other jurisdictions.

The legal landscape

Criminal proceedings against MRF have already occurred in other countries. ⁹⁷ According to an unpublished census presented at the 2017 World Conference on Research Integrity, 39 researchers from seven countries have been subject to criminal sanctions between 1979 and 2015. Penalties ranged from suspended sentences to 15 years in prison. ⁹⁸ Lately in Mainland China, capital punishment has also been proposed to punish fraudulent researchers. ⁹⁹ The US has a particularly long history of such proceedings ¹⁰⁰ brought under a variety of offences contained within the False Claim Act, ¹⁰¹ the federal conspiracy statute, ¹⁰² the false statement statute ¹⁰³ and the federal mail or wire fraud statutes. ¹⁰⁴ In Australia, two researchers have been convicted under the Crime and Misconduct Act 2001 in 2016. ¹⁰⁵

The most relevant legal instrument in England and Wales is the Fraud Act 2006. It came into force on 15 January 2007 and replaced the eight deception-based offences under the Theft Acts 1968 and 1978. The 2006 Act introduced a new general offence of fraud that can be committed in three ways by ¹⁰⁶ : (i) false representation (section 2), (ii) failure to disclose information (section 3) and (iii) abuse of position (section 4). There are two ancillary offences of ‘possession of articles for use in fraud’ (section 6) and ‘making or supplying articles for use in fraud’ (section 7). There is a maximum penalty of 10 years’ imprisonment on conviction on indictment.

In principle, MRF may also lead to legal actions under the consumer ¹⁰⁷ and sale of goods legislation, ¹⁰⁸ and cases that involve commercial interference, conflict of interests and Internet-based activities could invoke other offences. ¹⁰⁹ There are specific rules that govern research on human tissues, ¹¹⁰ drugs, ¹¹¹ and medical devices, ¹¹² violations of which can lead to a fine or imprisonment. In Scotland, criminal fraud is dealt with under the common law offences of fraud, uttering and embezzlement. ¹¹³ A number of statutory offences are available such as the Criminal Justice and Licensing (Scotland) Act 2010 ¹¹⁴ and the Computer Misuse Act 1990. None of the above, however, has been tested in the context of MRF within the UK.

Related case law is also limited. The first historical case of MRF in the UK involved James Shearer, an American physician working in the British Army as a sergeant in 1910s. He was tried by court martial and sentenced to death by firing squad. (The sentence was commuted to penal servitude.) ¹¹⁵ An isolated case in modern times involved Steven Eaton, a bioanalyst, who was assessing new drugs to be used in clinical trials. ¹¹⁶ He was successfully prosecuted under the Good Laboratory Practice (GLP) Regulations 1999 that regulate activities at test facilities undertaking health and environmental safety studies. ¹¹⁷ Section 7(1) of the Regulation requires adherence to ‘the conditions under which the study is planned, performed, monitored, recorded, archived and reported’, while Eaton had manipulated data to look as though trials had been successful, when they had in fact failed. However, the Regulations do not apply to facilities that are not member of the GLP programme, and the imposable penalties are arguably light; Steven Eaton was jailed for 3 months.

Little else has been written about how MRF is to be handled by the criminal law in the UK. Unlike in the US, where a wealth of literature exists on the subject matter, there is as yet no recognized or suggested legal framework within which MRF can be treated. It has not been a topic for discussion by those who support criminalization, and although a number of relevant legal instruments are available, there is little guidance on how to locate an applicable offence and to apply the law in a predictable and consistent manner. The threshold for prosecution is likewise unclear. An appropriate legal framework is essential if there is to be an informed and meaningful discourse on the subject matter. The following sections will examine the characteristics of the author’s proposed framework and explore how available legal instruments may be applied to achieve the desired purposes.

Scope of the proposed legal framework

At issue here is whether MRF should be treated as a specific and stand-alone offence or be subsumed under a general offence of fraud. Creating a specific offence of MRF demonstrates commitment on the part of the government to addressing society’s imminent concerns. The threshold for prosecution and penalties can be tailored to reflect the severity of different types of research misconduct and society’s disapproval of them. It would send a clear message to medical researchers that the quality of their work has societal impact, and that fraudulent behaviours will not be tolerated. The main difficulty with this approach lies in the need to decide what will constitute ‘medical research’ for the purpose of prosecuting a particular act or violation of required standards. ¹¹⁹

The term ‘medical research’ covers a wide range of activities from laboratory, bioinformatics and epidemiological studies to clinical trials on drugs and medical devices, all of which are subject to misconduct on the part of the researchers. What will amount to ‘medically related’ research is contestable since many inquiries in science or even the humanities could have implications for health and healthcare. Researches in engineering, for example, can impact public health and safety; economics analyses might influence healthcare funding. Questions may be raised about whether qualifications for the term ‘medical’ should be determined by the offender’s professional status, employment or institution; or by the nature of his work, its publication, application or academic field; or by the nature of its consequences such as medical injury or research grant fraud.

The term ‘research’ is also variably defined. In the ordinary sense of the word, research can mean:

a studious inquiry or examination; especially investigation or experimentation aimed at the discovery and interpretation of facts, revision of accepted theories or laws in the light of new facts, or practical application of such new or revised theories or laws. ¹²⁰

An alternative approach would be to pursue MRF under a general offence of fraud. This has the advantage of eliminating many of the definitional and technical issues associated with a specific offence. Its ‘catch-all’ effect would cover all research activities across all disciplines; in fact, the activity would not have to involve primary research at all (e.g. Steven Eaton’s testing of new drugs). This approach is also commensurate with the present trend within the UK to move away from over-particularized offences, ¹²⁵ a relevant example of which are the, now defunct, deception-based offences contained within the Theft Acts 1968 and 1978 that had been criticized for being overlapping and too technical. ¹²⁶ These offences required the prosecution to identify a victim ¹²⁷ and to prove that the victim would not have acted as he did but for the deception. ¹²⁸ These requirements have proved to be problematic, ¹²⁹ and a specific offence of MRF is likely to be plagued by similar problems. By contrast, the single general offence provided by the Fraud Act 2006 focuses on the conduct of the accused and is not attached to any specific activities. ¹³⁰ As will be further discussed, for a charge brought under the 2006 Act to succeed, it will be sufficient to establish that the accused has acted dishonestly when intending to make a gain, cause a loss or the risk of a loss; no actual gain or loss needs to occur. ¹³¹ All these are important considerations in MRF where gain or loss and its causation can be difficult to establish. The author favours this approach.

Guilty acts

The next task is to identify research behaviours that should attract the attention of the criminal law – essentially establishing the actus reus of the offence. Currently, there is no universally accepted taxonomy of wrongdoing in medical research. ¹³² The definitions of research fraud or research misconduct vary among regulatory agencies across European countries ¹³³ and within the US. ¹³⁴ The European Medical Research Council defines misconduct as ‘broadly falling into three categories: piracy, plagiarism or fraud’. ¹³⁵ The RCUK defines misconduct as ‘fabrication, falsification, plagiarism, or deception in proposing, carrying out or reporting results of research and deliberate, dangerous, or negligent deviations from accepted practice in carrying out research’. ¹³⁶ The UKRIO’s definitions include (i) fabrication, (ii) falsification, (iii) misrepresentation of data and/or interests and/or involvement, (iv) plagiarism and others. ¹³⁷ In academic literature, the term ‘research fraud’ customarily refers to the triad of FFP. ¹³⁸ This article will focus on fabrication and falsification.

There is no official definition of fabrication or falsification. Fabrication refers to the making up of false data without actually doing any experiment or study. Notable examples include Robert Slutsky, who published studies on cardiology that never took place, ¹³⁹ and Malcolm Pearce from the UK. ¹⁴⁰ Falsification refers to the manipulation of research data or processes or omitting critical data or results; research takes place at least to a certain extent.

From a legal point of view, the differentiation between fabrication and falsification is probably not important; what matters is the specific impact or seriousness of an individual act. Fabricating an entire clinical trial, for instance, has more significant impact than, say, the making up of one of 500 research subjects. Similarly, the manipulation of a picture to illustrate a genuine observation is not in the same league as falsifying a drug’s safety profile. There is no readily available tool to measure impact or seriousness in this regard, but inferences can be drawn from instruments that govern the conducting of clinical trials.

Under the European Medical Agency’s ‘Guideline for the notification of serious breaches of Regulation (EU) No. 536/2014 or the clinical trial protocol’, a serious protocol breach is, among others, a breach that is likely to ‘affect to a significant degree’ the ‘reliability and robustness of the data generated in the clinical trial’. The assessment of these criteria depends on a variety of factors including the effect of a protocol breach on study design, the type and extent of the data affected, the overall contribution of the affected data to key analysis parameters, and the impact of excluding those data from the analysis. ¹⁴¹ This author proposes that these factors should be used to determine whether a case of fabrication or falsification would amount to a ‘guilty act’ of MRF. For example, in the case of a clinical trial of a new drug, the actus reus may involve the making up of research subjects that never existed, the failure to record or report the drug’s harmful effects, and/or an inflated account of the drug’s benefits. Other factors may be included and assessment will depend on a finding of fact on a case-by-case basis.

Mens rea requirement

It is an established doctrine in common law that there shall be no conviction unless the accused have acted intentionally or recklessly, depending on the criteria of each crime. ¹⁴² This article will not venture into the complex issue regarding the mens rea requirement for fraud, but several issues need to be addressed. ¹⁴³

First, attributing culpability in MRF based on mental state concepts can be problematic. Words like ‘intentionally’, ‘knowingly’, ‘recklessly’ and ‘negligently’ have been used to describe an offender’s state of mind in MRF but the separation from general incompetence or mere ‘sloppiness’ may not be as easy as it might seem. ¹⁴⁴ In research, hypotheses are constantly made and tested. Errors may occur during experiments some of which may go undetected despite the best effort and intention. The selective presentation of relevant facts is also part and parcel of an accepted practice in scientific writing. Difficulties may arise when assigning culpability to a researcher who had failed, say, to report important data without any actual intention to mislead. Neither would it be easy to prove beyond a reasonable doubt that a researcher has intended to falsify data where complex statistical analyses had been involved. Culpability in cases of fabrication, on the other hand, is probably more straightforward to establish.

Second, motives are irrelevant. Though not legally binding, the RCUK Policy and Guidelines on Governance of Good Research Conduct requires that researchers ‘must not’ commit fabrication and falsification. ¹⁴⁵ Vanity or a well-meaning desire to ‘correct the literature’ will not provide a defence; it is the intention to publish faked data that matters. ¹⁴⁶ However, the offender’s motives may become relevant to the determination of penalties.

Third, although the concept of liability in criminal negligence is largely reserved for medical manslaughter cases, ¹⁴⁷ it can be applied to other areas. ¹⁴⁸ Conceptually, a researcher who had been grossly negligent in supervising the publication of a fabricated study that eventually led to a participant’s death could be held liable. ¹⁴⁹ A successful conviction would necessitate the establishment of a duty of care between the researcher and the affected parties, the introduction of standards of care in research, and the proof of a causal link from the supervisor’s conduct to subsequent medical treatment and patient injury. ¹⁵⁰ This is a tall order and an untested area involving novel legal paradigms that, in the author’s opinion, is unlikely to be visited under the prevailing climate.

Threshold for prosecution

An appropriate threshold for prosecution is essential if the law is to punish investigators such as Banerjee ¹⁵¹ or Wakefield ¹⁵² while leaving alone a postgraduate research student who falsified a dissertation. So far, the author has proposed that an act of fabrication or falsification that significantly affects the reliability and robustness of research findings could be prosecuted under a general offence of fraud. It will be further argued that only some, but not all, of these acts should be prosecuted, and that the threshold for prosecution can be determined by a test modelled upon the CPS policy on assisted suicide. ¹⁵³

Unlike other fraudulent activities that commonly result in identifiable and discrete quanta of property losses, an act of MRF can have multiple consequences of various kinds as described previously in this article. Some consequences are personal, while others are of public concern. They can be tangible like financial losses or intangible ones such as reputational damages and distrust. Importantly, harm may already have materialized or is simply anticipated. From a legal perspective, personal losses sustained by the offender are irrelevant. Materialized consequences with tangible and significant implications for the public (e.g. medical injury, misuse of government research grant) are the most serious. Merely anticipated consequences with minimal societal impact (e.g. a falsified student dissertation) would sit at the lower end of this spectrum of seriousness. Lying in between the two is an obvious mixture, the significance of which can only be assessed on a case-by-case basis. The task at hand is therefore to identify those cases that should attract criminal penalties – an issue that is not totally dissimilar to that of deciding whether or not a prosecution is deemed justifiable in assisted suicide cases. In that connection, the CPS policy on the latter offers some useful guidance. ¹⁵⁴

It will be remembered that the CPS policy on assisted suicide cases was the result of the House of Lords decision in Purdy that required the Director of Public Prosecutors ‘to clarify what his position is as to the factors that he regards as relevant for and against prosecution’. ¹⁵⁵ Under the Two-stage Full Code Test, once there is sufficient evidence to justify a prosecution (the Evidential Stage), prosecutors must go on to consider whether a prosecution is required in the public interest by referring to the general factors set out in the Code for Crown Prosecutors as well as lists of specific factors tending for or against a prosecution (the Public Interest Stage). ¹⁵⁶ A prosecution is pursued when the Two-stage test is satisfied.

A similar approach can be adopted for the treatment of MRF, and a list of MRI-specific public interest factors tending towards a prosecution is submitted herein:

The research has caused, or was likely to have caused, actual medical injury.

A list of factors tending against a prosecution may include:

There was minimal or no risk of the research causing medical injury.

These two lists are used to illustrate the overall approach and are neither exhaustive nor definitive. The operative principles would be the same as for assisted suicide cases and quoted herein:

Assessing the public interest is not simply a matter of adding up the number of factors on each side and seeing which side has the greater number. Each case must be considered on its own facts and on its own merits. Prosecutors must decide the importance of each public interest factor in the circumstances of each case and go on to make an overall assessment. It is quite possible that one factor alone may outweigh a number of other factors which tend in the opposite direction.

The absence of a factor does not necessarily mean that it should be taken as a factor tending in the opposite direction. ¹⁵⁷

Note that the CPS policy on assisted suicide provides that:

A prosecution will usually take place unless there are public interest factors tending against prosecution which outweigh those tending in favour. ¹⁶⁰

Putting it together

The authors’ proposed legal framework can be summarized as follows:

MRF is to be prosecuted under a general offence of fraud;

With this model, an act of MRF that poses a significant threat to the public interests would justify a prosecution and possibly a conviction with a heavy penalty (i.e. ‘serious offence of MRF’ – darkened area, Figure 1). A similar act with less significant impact on the public interest could be prosecuted and convicted but with a lighter sentence imposed (i.e. ‘lesser offence of MRF’ – chequered area, Figure 1). All other acts will not justify a prosecution (whitened area, Figure 1) with the possible exception of those committed by gross negligence (light grey area, Figure 1). The latter is an area of legal uncertainty as already alluded to.

OPEN IN VIEWER

Figure 1.

A schematic diagram showing different treatments of MRF activities. MRF: medical research fraud.

This model is not meant to offer a definitive solution. Due to the random nature of MRF, the assignment of blameworthiness may be affected by circumstances beyond the actor’s control, making it difficult to establish mens rea in some cases. For instance, it can be difficult to attribute consequences to the behaviour of the researcher where there are intervening factors such as commercial influences or institutional politics. ‘Lab lapses’ and sloppiness may also affect research findings without any ill intention on the researcher’s part. As such, the use of ‘seriousness’ alone as a guide for sanction assignment can be problematic, especially when the determination of seriousness would depend on the set of ethical appeals employed. ¹⁶¹ The problem may be (partially) remedied by introducing the concepts of foreseeability and role-related duties although their applications in the context of MRF can only be speculated. Furthermore, the assessment of an act’s impact on research findings and its consequences would ultimately depend on expert witness opinions, which can be difficult to locate in the unvisited realm of MRF. Notwithstanding, this model does provide a systematic approach to answering how MRF may be criminalized if needs be. Next, the author will examine whether this legal framework would sit well with the most relevant legal instrument in England and Wales – the Fraud Act 2006.

Fraud by false presentation (section 2)

This is the broadest offence under the Act. The actus reus is complete as long as a person made an ‘untrue or misleading’ presentation ¹⁶² : it is irrelevant whether the false representation is believed or has any effect on any other person. ¹⁶³ A representation may be express or implied, in words or communicated by conduct. ¹⁶⁴ The question of ‘dishonesty’ is one for the jury and the two-stage Ghosh test is applied. ¹⁶⁵ Stage one is an objective test that asks whether a defendant’s behaviour would be regarded as dishonest by the ordinary standards of reasonable and honest people. If answered in the positive, the second-stage subjective test will ask whether the defendant was aware that his conduct was dishonest and would be regarded as dishonest by reasonable and honest people. The prosecution only needs to prove that a false representation was communicated by the defendant to a person or to a machine. ¹⁶⁶

It can be readily observed that section 2 will cover research activities across all disciplines. Fabrication and falsification, however defined, constitute untrue or misleading representations and will almost certainly be caught. For example, it will not matter whether a researcher actually obtained a research grant; it will be sufficient to prove that he had included fraudulent data in his application with the intention to secure the grant. Since the offence covers representations made to a machine, an author cannot claim innocence on the ground that he merely submitted the manuscript to an editorial office’s computer. The absence of a requirement of actual harm also means that Wakefield could be found liable even if his study has not materially caused any rise in MMR infection rate or death rate. ¹⁶⁷

The merits and demerits of the Ghosh test have attracted scholarly attention. ¹⁶⁸ Of particular concern in the present context is that, when applying the stage two subjective test, it can be difficult for the jury to decide whether the researcher (e.g. a molecular biologist) would have been aware of what a layperson’s idea of proper research conduct would be. However, it must be noted that following the Supreme Court judgment in Ivey, ¹⁶⁹ the objective Barlow Clowes Test for dishonesty will likely apply whereby it will be irrelevant that the defendant had subjectively thought his conduct was honest. ¹⁷⁰

Another potential problem concerns section 2(4) that effectively covers most, if not all, forms of communication in research including journal publications, conference presentations, media interviews, course-works or even private discussions with peers. The fact that a representation can be ‘implied’ may subject an author’s opinion to unpredictable and possibly erroneous interpretations as to whether it is untrue or misleading. Section 5(2)(b) provides that gain or loss extends to not only gain or loss in money but also ‘any property whether real of personal (including things in action and other intangible property)’. The term ‘intangible property’ is not defined, however, and, in the author’s opinion, could be interpreted to cover non-financial gain such as prestige.

Authors are now required to declare conflict of interests when submitting a manuscript, but whether a failure to do so accurately warrants criminal sanction under section 2 is unclear. In January 2008, a woman submitted false information on her son’s school application form with a view to securing a school place. The school council commenced proceedings but later dropped the case. ¹⁷¹ The then Secretary of State of Schools, upon legal advice, opined that the Fraud Act 2006 did not apply in those circumstances. ¹⁷² However, in early 2017, a woman who faked her identity and invented dependent children to obtain university nursing places and NHS bursaries was given a suspended prison sentence. ¹⁷³ Further official guidance on the application of this widely drafted provision, and the meaning of ‘gain’ and ‘loss’ in the context of MRF, is needed.

Failure to disclose information (section 3)

This offence holds a defendant liable for failing to disclose information ‘which he is under a legal duty to disclose’. There is no definition of the term ‘legal duty’, and its existence is a question for the jury. ¹⁷⁴ The CPS commented that a legal duty may arise from statute, a contract, customs of a particular trade, a fiduciary relationship and others. Presumably, this can be extended to cover duties assumed by a grant applicant, the Principal Investigator of a clinical trial, or a scientist performing contracted research. Whether a legal duty exists between an author and his publisher for the purpose of section 3 is less clear. This offence will likely be invoked where research data were removed without disclosure; the defence is one of strict liability as the defendant cannot plead ignorance of the existence of a legal duty.

As with section 2, section 3 can potentially identify liabilities in a wide range of research behaviours, some of which are not normally considered as fraudulent. Since there is no requirement that the undisclosed information must be ‘material’ or ‘relevant’, a decision not to report data that have no bearing on the overall research findings can theoretically complete the actus reus. The defendant will then have to rely on the absence of dishonesty in which case it will be up to the jury to decide if the practice is indeed acceptable. The importance of defining the ‘guilty acts’, as alluded to in previous sections of this article, is key.

Abuse of position (section 4)

This offence concerns a person occupying a position in which he is expected to safeguard, or not to act against, the financial interests of another person. Again, the terms ‘abuse’, ‘position’ and ‘financial interests’ are not defined, and uncertainties such as those found with section 3 may arise. The CPS describes ‘position’ as ‘a position of trust’ that could include business partners or employees. The intended loss must be financial here which means research grant applicants could be implicated. ‘Abuse’ may include the conduct of an omission so that a laboratory supervisor who has forgotten to crosscheck research data could find himself liable. Again, this offence can potentially catch many types of behaviour in medical research that are not usually deemed blameworthy by ordinary societal standards.

Ancillary offences

Section 6 (possession etc. of articles for use in frauds) was designed to address developments in information technology and their impact on the ease with which a fraud can be committed. ¹⁷⁵ Section 7 (making or supplying articles for use in frauds) provides that a person is guilty of an offence if he knowingly makes, adapts and supplies articles for use in frauds. The inclusion of these provisions is timely in light of the emergence of open-access publication platform and cybercrime in medical research. For the purposes of sections 6 and 7, an ‘article’ is defined extremely broadly and can potentially cover anything from laboratory instruments to computers. The term ‘control’ finds its meaning in common law so that an article is in a person’s control if it is accessible to him. ¹⁷⁶ Therefore, a drug company’s ‘ghostwriter’ cannot defend himself by claiming that he was not the one who ordered the manipulation of data, whereas the broad definition of the offence also means that an honest scientist with an unpublished manuscript in his computer could, in theory, be found liable.

A conspiracy to defraud

The 2006 Act retains, amidst controversies, the common law offence of a conspiracy to defraud. This is not the place to discuss the pros and cons of this offence but while substantive offences should be used wherever possible, ¹⁷⁷ there are indeed complex cases that could be more effectively prosecuted as a conspiracy to defraud. ¹⁷⁸ Examples might include cases that involve international commercial sponsors, multiple researchers, institutions and victims. The proper prosecution of such cases under the substantive offences can potentially lead to a large number of counts and separate trials. By contrast, a charge of a conspiracy to defraud can facilitate a clearer presentation of the overall picture and of the gravity of the offence. ¹⁷⁹ Moreover, there are cases that can only be prosecuted as a conspiracy to defraud such as where the accused cannot be proved to have had the necessary degree of knowledge of the substantive offence to be perpetrated. ¹⁸⁰ This could be applicable to, say, cross-border allegations.

In summary, the 2006 Act can be applied effectively to the prosecution of MRF. It provides a general offence of fraud that is unburdened by technicalities and definitional issues, and it is commensurate with the author’s proposed legal framework. The actus reus is that of a dishonest act that will cover most if not all cases of fabrication and falsification. There is no requirement to prove actual loss, which is often difficult to establish in MRF. Section 3 of the Act covers acts of non-disclosure that is an important aspect of MRF where data suppression can lead to serious medical harm. The general offence can affect those in positions of power whose works may be influenced by commercial interests. The two ancillary offences can target technology-based cases, their perpetrators as well as accomplices. The retention of the common law offence of a conspiracy to defraud can conceivably serve useful functions in the prosecution of complex MRF cases. The imposable penalties are also more severe than those given by other means.

The broadness of these offences can also be their weaknesses, however. The 2006 Act casts a very wide net that can potentially incriminate many ‘minor’ cases of MRF. The CPS advises that:

Prosecutors should be acutely aware of the public interest in such cases, bearing in mind the relative standing of the parties and pay particular regard to any explanation for the failure given by the Defendant. ¹⁸¹

It is a matter of identifying the appropriate threshold for prosecution and gauging the seriousness of consequences. Under the author’s proposed framework, prosecutions would be limited to certain ‘guilty acts’ of falsification or fabrication. Its Two-stage Full Code Test model would provide a pragmatic approach to the determination of prosecutorial threshold in a transparent and accountable fashion. Decisions on penalties could be based on the suggested lists of MRF-specific public interest factors. However, the proposed framework does not lend itself readily to the assignment of culpability in certain situations. Cases such as those involving junior researchers or multi-institutional collaborations may continue to pose problems. Nonetheless, when applied in concert with the 2006 Act, the author’s proposed framework would enable the proper prosecution of serious cases of MRF that have been troubling the medical and scientific communities as well as the government and the general public in recent years.