Governance strategies for responding to alarming studies on the safety of GM crops

Abstract

Alarming studies about the safety of genetically modified organisms have been appearing since the first genetically modified (GM) crops were authorised for market release. The debates about these studies have a recurring pattern with similar players, arguments and strategies. More importantly, they are open-ended because those involved cannot agree on the significance and value of these studies or the consequences that should be attached to them. The roots of these disagreements lie deeper and connect to different views on science, society, nature and food production. Usually, governments respond in a procedural manner by referring the study to their scientific advisory bodies. Thus far, they conclude that there is no reason to revoke or postpone GM crops authorisations. However, the debate continues with undiminished vigour. The current response strategies to alarming studies seem to actually worsen the status quo in the GM debate because the underlying disagreements are not addressed. Starting from the current governance structure, we investigate how European governments responded to alarming studies in the past and look into the consequences of these strategies on the course of the debate in order to identify lessons and pointers for the future.

Keywords:

• GMO
• biosafety
• alarming study
• technology
• governance

1. Introduction

Alarming studies initially highlight a specific technological risk that is possibly being disregarded in the existing governance system. Governments therefore usually address the issue as a structured and case-specific scientific problem. Secondly however, these studies trigger a broader debate on the social and political elements of the technological debate. These issues cannot be addressed in the context of a response to the alarming study. In this paper we analyse government responses to cases of alarming studies to underpin our claim that alarming studies are unsuitable vehicles for a wider democratic debate about technology. We identify a series of recommendations on how to respond to alarming studies within the current governance structure, while our main and overarching point is that other platforms are needed to enable conversations about responsible innovation.

The dynamics of the debate ignited by alarming studies are not technology specific; there are strong similarities in the debate around different controversial technologies such as shale gas, nuclear energy, genetically modified (GM) crops or vaccination programmes. In this paper, we will use GM crops as our main subject.

The use of genetic modification in agricultural applications has been subject to strict safety regulations for a long time. These regulations originate from the scientific consensus that GM crops could potentially present risks to humans and the environment. The European regulations therefore focus on safety aspects, whereas other aspects (such as principal, ethical or socio-economic aspects) are considered an individual or market choice and are therefore secured by European labelling regulations for GM food and feed.¹ It should be noted that this article is written from within continental Europe and our analysis conclusions will therefore be more relevant in this context than outside of Europe.

Companies, universities and research institutes look into the potential risks and short- and long-term effects of genetically modified organisms (GMOs) on humans and the environment (Nicolia et al. 2014). The result of this research, whether published in scientific literature or not, forms the basis for the safety assessments by licensing authorities. Market approvals for GM crops are subject to an extensive environmental and food safety risk assessment before being approved or denied for commercial release (European Commission 2001, 2003). After market approval, monitoring is implemented as an additional safety net to identify any indirect or unexpected adverse effects on the environment. The dominant consensus in the biotech sector and regulatory community is that the current risk assessment practice is adequate (or even too strict) and thus that GM crops that are approved for market release are safe. Among scientists, the nature and status of this consensus are recurrent topics of discussion (e.g. Hilbeck et al. 2015). To date there have been no incidents confirmed by governments or competent authorities in which GM crops approved for market release have caused direct harm to the environment or human health (Nicolia et al. 2014). This consensus is extended and reconfirmed by enlarging the body of evidence about the safety of GM crops through ongoing research.

1.1. Alarming studies reignite discussion

Although the vast majority of research indicates that no significant adverse effects can be expected from the use of GM crops approved for commercial release, studies concluding that certain (approved) GMOs do cause harm to human and animal health or the environment do crop up occasionally. We propose to call those ‘alarming studies’ because they alarm other scientists, risk assessors, governments, companies and the broader public of unknown risks that are possibly being disregarded in the existing system. Although these (scientific) publications seem to be directed at the scientific community, they are characterised by the fact that they address a broader public (and set of issues). This public evolves around the topic of GMOs but actively comes into ‘existence’ when an alarming study is published. The concatenation of texts on this subject over time has created a public (and counterpublic) with a specific discourse that is addressed by alarming studies (see also Warner 2002). Alarming studies provoke fierce debates and can lead to considerable disquiet and anxiety. More importantly, the debates are open-ended because those involved ultimately cannot agree on the significance of these studies or the consequences that should be attached to them.

Governments have a duty to safeguard human and environmental safety, which means that society expects them to respond to reports that one of these is at risk. In other words: governments must respond in one way or another to alarming studies. This is usually done by a reassessment of the study by scientific advisory bodies. However, the acceptance of the outcome of the assessment is severely hampered by the fact that the appearance of an alarming study will also reignite a broader discussion about the value of GM technology. Disagreements about these aspects will not and cannot be solved by reassessing an alarming study. Thus, underlying questions remain unanswered and the answers provided are not acknowledged, resulting in frustration for all stakeholders involved. As a result, the status quo in the GM debate seems to get worse.

1.2. Biotechnology debate characterised by multi-level disagreements

Despite the dominant consensus, not everyone agrees with the current governance structure for GM crops. The debate about the implementation of GM crop technology in society is characterised by multi-level disagreements about definitional, factual, scientific, interest-based, value-based, moral and metaphysical aspects (Bovenkerk 2012). Definitional, factual and scientific disagreements have been classified to be merely attributed to science, whereas interest-based, value, moral and worldview disagreements belong to the realm of politics.² In practice however, these elements are strongly intertwined in discussions about GM technology. We would like to add, perhaps redundantly, that scientific disagreements are usually not solely about facts and political disagreements do not strictly originate from values. The roots of these multi-level disagreements are based on differences in views on science, society, nature and food production. This has been illustrated amongst others by a report of the Netherlands Commission on Genetic Modification (COGEM 2013). The report illustrates that there is no general agreement about the facts or about the values regarding GM. Neither is there agreement about the goal this technology could serve or about the means that should be used otherwise to reach this goal.

For the purpose of this article, we integrate the broader means/ends separation with a commonly used distinction between four types of problems. It is based on agreement about the facts and values (structured problem), agreement about one of them (moderately structured problem – either technical or merely political) and agreement about neither (unstructured problem) (Hisschemöller 1993). Furthermore, depending on the type of problem, different solutions are proposed to best fit the characteristics of the dispute (Hisschemöller and Hoppe 1995) (Table 1).

Table 1. Characterisation of problems based on scientific and normative consensus.

	Values
		Consensus	No consensus
Facts	Consensus	Structured problem	Moderately structured problem
		Agreement about means and ends	Disagreement about ends (agreement about solution)
		Experts or bureaucrats	Consensus conferences
	No consensus	Moderately structured problem	Unstructured problem
		Disagreement about means (agreement about problem)	Disagreement about means and ends
		Expert committees	Learning strategy/public debate

Note: Adapted from: (Hisschemöller and Hoppe 1995) and (Bovenkerk 2012).

The rhetoric of the debate about GM crops should be understood as an unstructured problem (Mampuys and Brom forthcoming). From a European perspective, we will show that stakeholders persistently keep approaching the problem as either a structured problem or a moderately structured problem in which scientific experts or expert committees will remove uncertainties and provide pragmatic solutions that will lead to a definitive outcome. The reigniting discussion after the publication of so-called ‘alarming studies’ about the safety of GM crops proves over and over that this is not the case.

We think that further polarisation and escalation of the debate have to be prevented in order to give future discussions about the value of GM technology a chance. Therefore, we investigate potential improvements in responding to alarming studies from a governance perspective. In this article we look into the options for action (such as monitoring, initial response, adopting measures, selecting experts/advisory bodies, national and international alignment, follow-up actions and communication) open to governments and advisory bodies in response to the appearance of an alarming study. The options each have their own benefits, risks and pitfalls, from which lessons and pointers can be drawn. These are looked at in more detail and recommendations are made on how to respond to alarming studies with a focus on doing justice to both the facts and the values that play a role in the debate and the role of political decision makers vis-á-vis the societal political debate. With these recommendations, we aim to at least restrain the widening gap between the different frames of facts and values about GM technology.

2. Alarming studies

We propose to use the term ‘alarming study’ for a scientific or other study, the results of which may or may not have been published in a peer-reviewed journal, from which it can be concluded that a technological innovation (such as a GM crop) poses a threat to human and/or animal health and the environment. The term ‘alarming study’ refers only to the claim made and does not say anything about the veracity or otherwise of the study.

Alarming studies have been appearing since the first GM crops were authorised for placing on the market. We concluded before that the GM debate should be understood as an unstructured problem, leaving discussions about the value and position of alarming studies open-ended. Consequently, the names of these studies recur and accumulate in the debate whenever new ‘alarming studies’ make the news. In terms of Warner (2002), these studies address both a scientific and regulatory public as well as a broader (counter)public consisting of other stakeholders such as companies, NGOs, citizens and consumers. Alarming studies are therefore one of the most important triggers to reignite the GM debate and that is why we analyse them to assess the effect of governance response strategies on the GM debate. A key example of an alarming study is a paper by a French scientist about the health effects of the consumption of GM maize (Text box 1).

Text box 1: Example of an alarming study.

In September 2012, the Journal Food and Chemical Toxicology published the results of a study by Séralini et al. (2012) in which rats were fed GM maize and various concentrations of a glyphosate containing herbicide for a period of two years. Séralini and his team concluded that the rats fed with the GM maize and herbicide developed more and more serious tumours than the control group, and they developed the tumours earlier in their lives. Because the GM maize variety used in this research has been on the market for several years, the publication led to an international debate and several national authorities and scientific advisory bodies looked into the results to see if there was a need to revise their opinion that this GM maize was safe for consumption. They concluded that the study contained methodological shortcomings and that the conclusions could not be justified. Although the study was dismissed as being unsound and incorrect by advisory bodies and the European Food Safety Authority (EFSA 2012b), several NGOs and scientists disagreed and the debate in Europe continued. Eventually, the Séralini paper was retracted by the editor and later it was republished in another journal (RetractionWatch 2014). Despite the discussion and formal outcome, this study is moreover referred to as proof that the consumption of herbicide-tolerant GM maize causes cancer and other adverse effects (GMOSeralini 2014b).

Other examples of alarming studies that we analysed for the purpose of this article are Ewen and Pusztai (1999), Rosi-Marshall et al. (2007), Huber (2011) and Carman et al. (2013) (Table 2).

Table 2. Examples of alarming studies.

Author	Alarming effect claimed
Ewen and Pusztai (1999)	Rats fed GM potatoes show abnormalities in the digestive tract
Rosi-Marshall et al. (2007)	Harmful effects of GM maize in aquatic insects
Huber (2011)	New pathogen from GM crops
Séralini et al. (2012)	Rats fed GM crops develop cancer
Carman et al. (2013)	Pigs fed GM feed show harmful effects

The discussion following the publication of an alarming study appears to have a repetitive character that can roughly be described as follows: the author(s) and opponents of GMOs point to the alarming publication as proof that GM crops pose a risk and some argue that measures should be taken (immediately). Governments usually respond in a procedural manner by referring the study to their scientific advisory bodies. These assess the publication according to the agreed research protocols and methods based on EU and international regulations and come to the conclusion – so far – that the research does not pass the test of scientific credibility and that there is no reason to revoke or postpone authorisations for GM crops:

Taking into consideration Member States' assessments and the authors' answer to critics, EFSA finds that the study as reported by Séralini et al. is of insufficient scientific quality for safety assessments. EFSA concludes that the currently available evidence does not impact on the ongoing re-evaluation of glyphosate and does not call for the reopening of the safety evaluations of maize NK603 and its related stacks. EFSA statement on final review of Séralini et al. (EFSA 2012b)

Overall, the data presented in the paper are not convincing of adverse effects due to the GM diet and provide no grounds for revising FSANZ's conclusions about the safety of previously approved glyphosate-tolerant and insect-protected GM corn lines and glyphosate-tolerant GM soy lines. Australia New Zealand Food Standards' response to a feeding study in pigs by Carman et al. (FSANZ 2013)

For the government and those in the scientific community who endorse the current risk assessment frame, this appears to be the end of the matter. However, those who are of the opinion that the results are (another) reason to ban GM crops are disregarded and they continue to pursue the debate with undiminished vigour. Characteristic for this discussion on GMOs is the broadening of the debate to wider issues like the safety of GM crops in general, distrust in biotech companies, the power of multinationals over global food production and the desirability of current practice in agriculture (Marris 2001). Implicitly and explicitly, multi-level disagreements come to the surface. Some examples of arguments in this debate are:³

With regard to biodiversity, there have been too few long term, properly controlled studies of the effects of GM crops to make much comment. Rosi-Marshall case (Fitzsimons 2009).

Séralini's findings revealed that industry and regulatory claims of biological irrelevance of effects found in 90-day tests are invalid. They showed further that the regulatory system for GM foods is inadequate and cast into question the safety of all commercialized GM foods. Séralini case. (Robinson 2013)

This is highly sensitive information that could result in a collapse of US soy and corn export markets and significant disruption of domestic food and feed supplies. On the other hand, this new organism may already be responsible for significant harm. Huber case. (Huber 2011)

There are a lot of things that are wrong with GMOs, but not on a biological level. Rather, it's the question of intellectual property and patent laws that are a huge disadvantage to small farmers. Carman case. (Gorski 2013)

There is a real problem for us here, and that is that you say that it is not right to discuss unpublished work; as I understand, all of the evidence taken by the advisory committee in that report comes from the commercial companies, all of that is unpublished. Ewen and Pusztai case. (Williams 1999)

Similar (underlying) arguments are described by Sarewitz (2004) with regard to a ‘vicious’ debate that unfolded after the publication of Quist and Chapela (2001) about the introgression of transgenic DNA into traditional maize events in Mexico. This case could also be considered an alarming study. Under current policies, the debate on the authorisation of GM crops concentrates on demonstrating their safety. Although the existence of uncertainties in the environmental risk assessment is acknowledged, governments urge scientists and advisory bodies to come up with answers that can dispel these uncertainties (Van Asselt and Vos 2010; Wardekker et al. 2008). A standard model is that if uncertainties are removed, the correct course of action will be apparent (Sarewitz 2004). This suggests that governments assume that uncertainties about risks lie at the heart of the public debate on GM crops (Marris 2001), and it hopes that removing these uncertainties will settle the debate. Two issues can be identified in this approach.

Firstly, it disregards the diversity in scientific disciplines resulting in different and sometimes competing views on science and nature, as also demonstrated by Sarewitz (2004). According to Sarewitz, these differences reflect the ‘richness of nature, and the consequent incapacity of science ( … ) to develop a coherent, unified picture of “the environment” that all can agree on’. He concludes that ‘uncertainty is in part a manifestation of the disunity of science and the plurality of institutional and political players ( … ) involved in the conduct and interpretation of scientific research’ (Sarewitz 2004). Therefore, even ruling out ‘conflicts of interest and ideological commitments to look at “what the science is really telling” us can be a meaningless exercise’ (Sarewitz 2004).

Secondly, governments limit their role to ensuring public health, environmental safety and freedom of choice (e.g. labelling in the EU). This means that they hardly take non-safety issues into account in decision-making on GMOs, whereas the sentiment is largely concerned with just these issues. This changes the focus of the safety aspect, because scientific research and risk assessment themselves become the subjects of a critical political discussion in which broader arguments are intertwined with safety issues.

Given the recurring debate about GM crops, it seems that the current (risk) assessment framework and governance structure is controversial because it does not take into account broader issues that relate to environmental concerns and food production. We think that other platforms are needed to enable conversations about responsible innovation. However, a definitive solution that will settle all levels of disagreement does not seem close on the short term, if possible at all. Meanwhile, some practical solution has to be found within the current governance frameworks. Therefore, we identify a series of recommendations on how to respond to alarming studies within the current realm of technical debate about alarming studies.

3. Governance options

Within the force field of different arguments that are put forward in response to an alarming study, national governments are looked at for a solution. Given their responsibility towards those who rely on the existing framework such as consumers, but also companies, researchers and investors, governments have to respond to alarming studies the best they can. This is a challenge, because those involved have different ideas about the knowledge and actions necessary to solve the problem (Sarewitz 2004; Wynne 1989). Experiences from cases of alarming studies from the past can throw light on the effect of various options open to governments and advisory bodies as a learning opportunity for the future. The government's response will depend on various factors, such as the status of the research, the nature and intensity of the debate and the stage it is in.

In this article, we distinguish four phases (I–IV) in which the government and advisory bodies have an opportunity to make choices: preparation (monitoring), first response (timing and opening gambit), obtaining advice (gather expertise and data) and finally the phase of response (communication and follow-up actions) (Table 3). For each phase, we will discuss the possible consequences of government actions and we will identify pointers and pitfalls.

Table 3. Governance of alarming studies: phases and questions.

Phase	Policy questions
I	Preparation: monitoring	Monitor scientific developments? Monitor societal debate?
II	First response: timing and opening gambit	Respond or ignore? Moratorium or other measures?
III	Obtain advice: expertise and data	Choice of advisory bodies? Coordination with European advisory bodies?
IV	Response: communication and follow-up actions	Attune communication to stakeholders? Follow-up actions: further research?

3.1. Monitoring (I)

Preparing for discussions about alarming studies is usually done by means of monitoring via civil servants, assessment agencies and advisory bodies. Through their work and (inter)national networks they are naturally informed of new developments and ongoing research. By monitoring scientific developments, governments can avoid being confronted unexpectedly with an alarming study making the news and they will be able to respond rapidly. Monitoring can also support a learning process on recognising emerging discussions. A possible pitfall for agencies interpreting monitoring results is that results that go against the dominant scientific consensus are not picked up as quickly after repeated false claims (by the same author). This makes it important to remain open to new developments and unconventional research methods. Another possibility is that researchers themselves inform governments about ‘alarming’ results at an early stage. Regardless of the veracity of alarming results, the course of the debate and monitoring thereof can also influence the decision and timing for a response from the government.

3.2. Timing and opening gambit (II)

After publication of an alarming study, governments have to decide about their (initial) response to the event. First, relevant questions have to be answered such as: What is the status of the publication and are immediate measures, such as a moratorium, justified? Answering these questions is difficult, especially when there are uncertainties about the scientific value of the publication. Ignoring an alarming study or responding too late can lead to suspicion and dissatisfaction with the conduct of the government, but an ad hoc response from governments invoking precautionary measures will give the impression that the situation must be serious, which can also lead to an (possibly unjustified) escalation of concern.

3.2.1. Initial response

Alarming studies can lead to national or international debates about the safety of GMOs, and in most cases it is not possible to determine straight away whether the results and claims in such studies are valid, regardless of whether the study is published in a peer-reviewed journal or not (Text box 2).

Text box 2: Peer review as a criterion for assessment.

Several alarming studies have put forward the question whether the government should respond to peer-reviewed studies only or to unpublished but alarming research results as well. Peer review is considered by both proponents and opponents of specific studies to be a hallmark of good research, but they draw different conclusions: proponents say the research was good, whereas opponents say the peer review process was flawed. Both the value and limitations of peer review have been profoundly discussed in literature (Bohannon 2013; Jennings 2006; Nature 2006; The Economist 2013). In addition, it should be noted that applications for marketing authorisation of GM crops can also contain unpublished and non-peer-reviewed information. Although these studies undergo a type of review in the form of appraisals by competent authorities and advisory bodies this could become a pitfall that makes the government come across as biased. Peer review seems therefore not a decisive argument whether or not to further investigate an alarming study.

In the already polarised debate about GMOs, alarming studies are very likely to make the news. With the chances of considerable media attention, the EFSA decided to respond instantly to the publication of the Séralini paper (EFSA 2012a). In a statement, EFSA acknowledged the existence of the study and announced an assessment of the paper's relevance. The fast response demonstrated an effective monitoring of scientific developments. On the other hand, to those who are less familiar with the regulatory framework it can also invoke questions about the working methods and function of the organisation such as: are there GMOs in our food and are they safe? Is this the first time that EFSA has looked into the safety of this maize?

Governments and governmental bodies usually limit their role to ensuring that certain requirements are met, such as safety standards and (specifically for Europe) freedom of choice with regard to food products by means of labelling. The underlying objectives and methods for decisions about these issues are often implicit or largely unknown to the public. This goes in particular for choices that have been made before alarming studies gather media attention, such as the regulatory approval of GM crops. If people are not aware of the existence of safety regulations for the commercialisation of GM crops, the acknowledging statement mentioned above may not have the contemplated effect. The decision-making process of GM crops in Europe has moreover been deemed indistinct and complicated by scientists, companies and (non-EU) regulators/policy-makers (Dunwell 2014; Masip et al. 2013). Besides a lack of transparency, these complexities have led to several problems in international trade (Meester, Berkhout, and Dries 2013). This likely makes it even more difficult for the broader public to get a grasp on who decides when about GM crops and based on what kind of data and considerations.

Studies about consumer attitudes to GM food support the idea that providing clear and convincing information about managing the risks of a technology is more effective than promoting the benefits (and thereby ignoring concern about potential risks) (Durant and Legge Jr 2006; Hess et al. 2013). The availability of transparent and understandable information about procedures and regulations could therefore be helpful to answer relevant basic questions the moment an alarming study hits the news. Obviously, the more complex or political procedures are, the more difficult it will be to translate this information outside a group of experts involved in the regulatory framework itself.

3.2.2. Moratorium

Based on the results of an alarming study, governments have to decide whether or not immediate (precautionary) measures should be taken, such as a moratorium on the import and/or cultivation of a GM crop or even retrieving products from the supermarket shelves. Changing existing procedures or declaring a moratorium are only possible if there is a justifiable ground for concern regarding human or environmental safety. However, in the early stages of a study that goes against the dominant regulatory consensus, there are usually several uncertainties about the seriousness of the situation. Given the body of evidence about the safety of GMOs that have been approved for market release, a single alarming study or publication will seldom provide indisputable evidence of an acute risk that justifies immediate action. Additionally, a moratorium can have far-reaching consequences for political relations and commerce. For these reasons, the various interests at stake and the risks involved have to be carefully considered before taking restricting measures.

However, it should be noted that in risk perception, the seriousness of the effect is often considered of greater importance than its probability. While this can be a sign of irrationality (‘probability neglect’ Sustein 2005), it may be morally justified. For example in the case of a catastrophic risk such as a nuclear meltdown, the consequences can be enormous, even though the chance of occurrence is low. This can justify that people find it an unacceptable risk (Roeser 2002); also given that risk estimates involve uncertainties (Hansson 2004; Slovic 2000). Feelings of concern regarding food safety can incur emotions that are likely to be magnified by media releases with shocking images, such as photos of rats with enormous tumours as an acclaimed result of GM food consumption (photos from the Séralini et al. (2012) paper were moreover reproduced in news articles e.g. Castillo (2012)). Emphasising the safety of a GM crop and pointing towards the existing safety assessments for market approval are unlikely to contribute to the acknowledgement of these emotions.

Besides national or European measures such as a moratorium, there are other ways to enable people to act according to feelings of concern in relation to GMOs. In the EU, GM labelling in food provides a means for people to choose whether to consume GMOs. The effectiveness of these measures lies in the availability of transparent and understandable information about regulations such as those for labelling and traceability (Paragraph 3.2.1).

3.3. Expertise and data (III)

In most cases, the government will only be able to make a decision about any measures to be taken after establishing with sufficient certainty if the study justifies these actions. In the current assessment framework of GM crops, this is usually done by having the study reviewed by scientific advisory bodies.

3.3.1. Choice of advisory bodies

Various (inter)national advisory bodies are appointed to assess the environmental and food safety aspects of GMOs and provide advice to their government. Therefore, it seems evident to ask them for a reassessment. However, asking the same bodies that have assessed the GM crop in the first place for advice on an alarming study about that same crop can incur resistance because of the possibility of bias or prejudice (GMWatch 2012a). This is one of the reasons why the Belgian Biosafety Council chose to install an ad hoc advisory committee to examine the Séralini study (Belgian Biosafety Advisory Council 2012). The decision to either lay down the request with the appointed advisory body or install an ad hoc committee each has its own advantages and disadvantages.

Appointed advisory bodies have relevant expertise and experience and the resources to respond rapidly. Additionally, their visibility and ‘raison d'être’ are enhanced in the process. However, the provision of advice by the established advisory bodies also has pitfalls. The ability to remain sensitive and alert to new research methods and/or results may be constrained, because established bodies usually work with specific research protocols and methods, based on EU and international regulations. Moreover, in the interests of consistency and reliability, professional organisations may be reluctant to retract earlier advice, given that their reputation and credibility are important factors in how their advice is regarded. Scientific knowledge is subject to change, whereas a certain degree of consistency is also important in risk assessment and evaluation. Constantly changing the risk assessment can lead to a loss of confidence among companies (legal certainty) and other stakeholders, whereas waiting too long to make changes can also damage confidence. Furthermore, when advisory bodies become (or are seen to be) part of the established system that represents a dominant but disputed paradigm, their advice may not provide useful input, and may even exacerbate the controversy (GM-FreeCymru 2012; GMWatch 2012b).

Obtaining advice from a different or ad hoc advisory body can provide a fresh perspective and may be more easily accepted. However, by establishing an ad hoc committee the government also risks putting the value of the acknowledged advisory bodies in doubt. It could even disqualify them in the eye of the public. Furthermore, putting together an ad hoc committee is time-consuming and there is a risk that the government will be accused of cherry picking from the pool of available expertise to influence the outcome. The decision to consult an established advisory body or an ad hoc advisory body is namely a political one.

3.3.2. Coordination with national and European advisory bodies

Where food and environmental issues surrounding GMOs have international ramifications it is customary to consult with other (inter)national advisory bodies about the quality and availability of the research data (Text box 3) and the validity of the conclusions (EFSA 2012b; NVWA 2012).

Text box 3: Availability of research data as a criterion for assessment.

The availability of (raw) research data of alarming studies as well as safety assessment data from biotechnology companies is a recurring topic in the GMO discussion. While we will not elaborate on this point in detail here, we want to point at the limitations of the persuasive power of data. Interpreting third-party raw data is difficult and time-consuming and if, after all the hold-ups and delays to provide this data, this does not lead to a definitive judgement about an alarming study, stakeholders may not be forgiving (given the sense of urgency and concern). Furthermore, the data debate suggests that providing missing information will lead to a final solution and that there is a single truth waiting to be uncovered. It should be noted that the disagreement about GMOs is not solely about the data themselves but merely about the interpretation and relevance of these data (and who provides it).

The advantage of (international) consultation is that use can be made of a wide variety of available expertise. If it is possible to come to a common position, coordination has the potential to achieve international agreement, which can in turn increase acceptance and support for the outcome.

A risk of (international) coordination is that the agreements made will look like consensus while they are in fact a compromise. Given the different positions within Europe about the authorisation of GM crops, some points will be subject to negotiation: the evaluation and measures proposed may be too cautious for some and too rigorous for others. Presenting compromises in a consensus document can be conceived as a selective form of transparency in which minority opinions are not visible. Furthermore, there is a risk of tunnel vision among those in agreement. A risk of limiting consultations to national coordination is that the final outcome may not be in agreement with international (EU) policy or that different outcomes fuel the GM debate leading to further polarisation.

To the broader public, international coordination can be seen as an obscure process, making it easy for critics to dismiss it as a conspiracy or orchestrated response (GM-FreeCymru 2012). In addition, it may increase a defensive position of the authors of the paper and decrease the possibility to have an open and fruitful discussion. The integrity and authority of the negotiators is therefore of importance during international coordination. As borders fade, international orientation and coordination are often essential (or even obligatory within the EU), but they do not relieve national governments of the duty to explicitly form their own judgement.

3.4. Communication and follow-up actions (IV)

Once the assessments by scientific advisory bodies are carried out, governments have to decide on the value of the alarming study and any consequences that should be attached to it. Sometimes, politics invoke scientific uncertainties as a reason to justify taking or not taking action in relation to controversial technologies when in fact underlying issues or values play a role (Sarewitz 2004). A recent article by Holbrook and Briggle (2014) also elaborates on the complicated relation between knowledge and action. Amongst others, they underline the importance of addressing or even confronting the underlying values as an essential part of responsible innovation and decision-making. This puts up the question whether and which other arguments apart from science the government should consider and what channels they should use to communicate their view on an alarming study.

3.4.1. Communicating about the assessment of alarming studies

In the first instance, alarming studies about the safety of GMOs have a bearing on safety policy. Governments respond by asking their advisory bodies to investigate whether the results of the study are accurate and valid. After a scientific assessment, they will have to make a statement why they will or will not take measures.

These statements usually reconfirm the dominant consensus about the safety of GM crops approved for market release. So far, it was concluded that there is no reason to revoke or postpone authorisations for GM crops based on the results of the alarming study. On the one hand, this response addresses the primary question about the initial trigger for the discussion: the alarming study. On the other hand, it does not clarify which arguments are considered in the formal decision-making on GMOs (e.g. biosafety) and which are not (wider issues) and why. Usually, communication is limited to pointing out the relevant procedures (‘we have asked our advisory bodies to assess or reconsider the study’) and issuing a press release on the findings of these advisory bodies. While this is relevant and important, it all too easily comes across as ‘these are the facts and you'll have to accept them’ (Te Molder 2011). It remains unclear why other arguments, for example in a broader debate on agriculture and food production, are not taken into account.

Contextual arguments regarding GMOs are part of the political debate about the type of society we want and the technologies we need to get there. They concern topics such as sustainability, biodiversity and naturalness, but also monopolisation of the food chain and the power of biotechnology companies. These arguments frustrate the dialogue because a solution to these wider problems cannot be found within the specific situation of an alarming study. Ad hoc situations, such as the appearance of an alarming study, are therefore by definition unsuitable vehicles for these arguments and discussions. Nevertheless, political arguments are out of necessity triggered by the appearance of alarming studies because they cannot be addressed elsewhere. Platforms that enable conversations about responsible innovation are urgently needed and the government can reinforce its position as a public's source of information by establishing a platform where political arguments regarding GMOs can be aired. However, the responsibility to establish these platforms is not confined to the government. There are different options for the form this platform could take, such as forums, ‘thinking laboratories’ (a cross between closed expert committees and open public debates) (Ruivenkamp and Jongerden 2010) or a broader (inter)national debate on food production in general. We emphasise that addressing wider issues involves more than setting up a discussion forum; it also obligates those involved, including the government, to take the outcome seriously and act on the results. Moreover, contextual arguments and wider issues are in fact a political discussion in which the role of public policy including regulation is at stake.

3.4.2. Follow-up actions

Alarming studies present new results that go against the dominant consensus on the safety of GMOs. The usual response to contradictory results is to establish whether they can be confirmed or refuted by existing studies or by conducting further experiments. However, if the experimental design and method differ from standard tests and broadly recognised research protocols, it proves difficult to find studies to compare results with (DeFrancesco 2013). Moreover, when the majority of scientists in the field consider a study to be methodologically flawed or incorrect, repeat studies will always lead to further discussion. Furthermore, from a scientific point of view a biased or methodologically dubious safety study (accidentally or on purpose) should not be an incentive to conduct further research with public sector funding (Text box 4). In such cases, the responsibility should lie with the researchers themselves to improve and rectify their study. This can be problematic because it is not always unequivocally clear whether a study is scientifically sound or not or whether the findings are unreliable due to misconduct or honest error (Hayes 2013).

Text box 4: Time and funding as a criterion for follow-up research.

Government funding for replicating a study branded by the scientific community as being methodologically weak can meet with incomprehension and resistance from the scientific field, Furthermore, counter-research into safety studies of GMOs found to be safe for placing on the market is not interesting for most researchers because of the high costs and limited publication potential. Additionally, it can be regarded as a waste of time, money and experimental animals. The compulsory post-market monitoring of GM crops should be sufficient to bring to light any unanticipated adverse effects following marketing authorisation. Furthermore, commercial activities may be put on the backburner as a precaution when follow-up research is initiated.

Undertaking further research confirms that the issue is taken seriously and it can bring an element of calm to the debate, at least temporarily. On the other hand, commissioning further research may also cause a public stir, because then there is ‘apparently’ reason for concern. Referring to existing and repeatedly confirmed research results is faster and can emphasise the existing body of evidence. However, because comparative research results are not abundant, there is a risk that the government and advisory bodies will be accused of cherry picking the results that suit them. It should be noted that study designs for complex research such as feeding trials differ to varying degrees. This will easily ignite another discussion about methods, execution and results.

If repeat or follow-up research is being undertaken, stakeholders will probably try to influence the design of the additional studies and the composition of the research team. More importantly, they may withdraw support during the process if they disagree with the chosen approach. These situations already occurred with follow-up research of the findings of Séralini et al. (2012), initiated by the French government (GMOSeralini 2014a) and the EU GRACE project.⁴ If involvement of important stakeholders is lost, support for the final outcome of the research will also crumble. Given the polarised discussion about GMOs, there is a significant chance that when the results of additional research are published, the debate will reignite and escalate all over, whatever the results (Bauer-Panskus and Then 2014; Schiemann 2014). Points to consider when referring to already existing research results include identifying clear criteria or a quality model for selecting alternative studies to refute or confirm the alarming study. Things that require careful attention when commissioning further research include the choice of researchers and the transparency of the decision-making process.

There seems to be no ‘right choice’ in conducting additional research or not. If no repeat studies are carried out and no comparable results can be found, the study will remain unique and, despite the criticisms, references to the results will remain. It may therefore be advisable to repeat studies to overcome the limitations of the previous one. Although this will not solve the issue indefinitely, the potential for scientific improvement remains open and communication between stakeholders is encouraged.⁵

4. Discussion and concluding remarks

The lack of and focus on consensus about the significance and consequences of alarming studies makes these events important catalysts in the recurring debate about GMOs. They reignite a broader discussion about GMOs where conflicts of interest, quality of scientific research, government responsibility, transparency of safety research, influence of biotechnology firms and monopolisation of farming practices play an important role. There are multi-level disagreements about the value and contribution of GM technology to food production.

With the analysis of government responses to alarming studies, we have shown that the scientific assessment is part of the regulatory and socio-political debate about the safety or desirability of GMOs but it does not cover all its aspects. We consider it important to emphasise that scientific research on the safety of GMOs including follow-up studies will never produce indisputable proof. In this context, we would like to acknowledge two authors who have paraphrased the role of science in an excellent way. According to Oreskes (2004), science can at best produce ‘a robust consensus based on a process of inquiry that allows for continued scrutiny, re-examination and revision’. Sarewitz (2006) describes science as an excess production of objectivity: ‘Science seeks to come to grips with the richness and complexity of nature through numerous disciplinary approaches, each of which gives factual, yet always incomplete, views of reality.’ It is ultimately up to politicians to decide at what point science provides a sufficient basis to make a decision and to accept responsibility for this (Sarewitz 2004).

In practice, the formal process determining decision-making about GM crops strongly focusses on depoliticising the issue at hand and treating it as a technical ‘scientific’ or structured ‘factual’ problem. There is an extensive framework for the (risk) assessment of GMOs in Europe, but it seems that despite broad agreement amongst part of the stakeholders, others disagree with this governance structure. The recurrence of alarming studies and current response strategies of EU governments lead to further polarisation and hardening of the GM debate.

From the recurrence of alarming studies, it becomes clear that the current framework is controversial, but a quick fix for the complex problem of GM technology is not nearby. Given the complexity and duration of the GMO debate, it seems clear that there is no ideal recipe or roadmap on how to handle alarming studies about the safety of GM crops. Therefore, solutions have to be found within the existing framework and governance structure. To limit the widening gap of the facts and values discussion about GM crops, governments should improve their response strategies. This article discusses the various options for governments and advisory bodies in responding to alarming studies about the safety of GMOs. Based on an analyses of the dynamics of past debates, lessons, pointers, risks and pitfalls are identified that can help the government prepare for and anticipate future situations (Tables 4–6).

Table 4. Governance of alarming studies: preparation.

Phase	Lessons and pointers
I	Preparation	• Monitor scientific developments • Remain alert and open to new or even unconventional research • Ensure access to information about decision-making GMOs

Table 5. Governance of alarming studies: response and advisory requests.

Phase	Lessons and pointers
II	First response	•Respond •Acknowledge situation and explain governmental priorities •Redirect to existing and understandable information sources
III	Advice	•Emphasise importance of openness towards research results •Establish criteria for a second opinion •Visualise individual positions in international alignment

Table 6. Governance of alarming studies: response and follow-up.

Phase	Lessons and pointers
IV	Response	•Acknowledge both specific and contextual aspects •Make explicit which issues are not addressed and why •Repeat and improve study to overcome limitations •Be realistic about the outcome

Our first conclusion is that alarming studies about the safety of GMOs will always make the news because the GM debate has never settled due to multi-level disagreements about facts and values. Scientific and political disagreements both play a role amongst the public, including the regulatory framework and among scientists themselves. To prevent unexpected confrontations with alarming studies, governments should monitor the developments in scientific research by means of their network of civil servants, assessment agencies and advisory bodies. Providing insight into the rules and regulations regarding safety of GMOs after an alarming study hits the news will not be embraced easily or even be regarded with suspicion. Therefore, it is important that transparent and understandable information about both the scientific and political procedures regarding decision-making of GMOs is already available. This kind of continuous transparency and disclosure gives structure to the debate and provides a better background to understand the actions taken or not taken as a consequence of an alarming study.

Secondly, because it is not possible to determine straight away whether the results and claims in alarming studies are valid or not, their credibility and validity will always have to be investigated. An instant response or statement that the government is aware of the situation and visualisation of their priorities in responding to the alarming study can contribute to transparency and trust. Additionally, governments could redirect questions to existing information sources about food and environmental safety of GMOs and the decision-making process.

Given the polarised discussion on GMOs and the strong pro-/contra atmosphere, advisory bodies must beware of rigid thinking and tunnel vision and keep an open mind towards new or unconventional research methods and results. Governments should be aware of both a potential bias of conventional advisory bodies as well as the risk of disqualifying them by installing an ad hoc commission by establishing criteria for a second opinion in addition to the advice from the formally responsible bodies. This should make sure that the second opinion adds to the credibility of the system instead of discrediting it. In addition, advisory bodies and governments should be aware of the effect of international alignment on the escalation and polarisation of the debate.

Third, a (re)assessment by scientific bodies of the study or the GMOs concerned is not sufficient to bring the debate to a satisfactory closure. In the process, we claim it would be advisable to make a distinction between the information and communication needs of the different stakeholders, which will help the government and advisory bodies to devise appropriate communication strategies for different stakeholders. This also means the public sector must have in-house expertise on science and science communication to ensure that it can meet the information needs of the various stakeholders. Furthermore, communication should not start when an alarming study is published, but beforehand. That is part of what responsible innovation entails in our opinion. Responsible innovation means integrating societal and ethical considerations in science and technology research and development (Van Oudheusden 2014). Therefore, governments should not only communicate the safety issues in their response. Such a procedural reaction does not do justice to the political discussion about the context of GMOs that is brought to the table repeatedly. Although an (unmanageable) broadening of the debate could hamper the chance of reaching a consensus on the scientific value of the alarming study, governments should acknowledge the position and role of both scientific and political aspects in decision-making.

Furthermore, although more research will unlikely solve the issue of alarming studies and GM technology indefinitely; repeating a study to overcome its limitations keeps the potential for scientific improvement open. Despite the different viewpoints, communication between stakeholders should be continuously encouraged and an effort should be made to prevent withdrawal of important stakeholders from the process.

We think that the reigniting discussion about alarming studies on GMOs is an indicator in a broader debate about agriculture and food production, given the fact that many contextual arguments are not GMO specific. Governments should therefore continue to monitor social trends and developments and remain cognisant of the contextual arguments about GMOs. Forcing the discussion into a safety debate format leads to frustration and dissatisfaction. For various reasons, some cannot accept the conclusion that GMOs authorised for market release are found to be safe by scientific advisory bodies. At the same time, governments and scientists become frustrated because the public will not accept ‘the facts’ about the safety of GMOs.

It is essential to indicate which arguments will be considered in the decision-making on GMOs and which will not. More importantly, it should be made clear where these other arguments may contribute to policy-making, for example in a broader debate on agriculture and food production. This could partially discharge the scientific debate from normative questions regarding the desirability of GMOs. The first steps towards such a system were already taken in 2010 when the European Commission announced plans to give individual member states the freedom to veto cultivation of GM crops in their country (European Commission 2010). In 2015, this resulted in an amendment of Directive 2001/18/EC aiming to combine a community authorisation system regarding safety with additional freedom for Member States to decide whether they wish to cultivate GM crops on their territory based on non-safety arguments.⁶

It is ultimately up to politicians to decide when scientific evidence is ‘good enough’ to make a decision and stand for it. With this, the debate should shift towards the issues that are really at hand in the GMO discussion. Aside from procedural decisions, we conclude that it is necessary to put the political debate back on the agenda with regard to unstructured problems such as GM crops. To cite Sarewitz and Pielke Jr (2000)

No longer able to hide behind scientific controversy, politics would have to engage in processes of persuasion, reframing, disaggregation, and devolution, to locate areas of value consensus, overlapping interests, or low-stakes options ( … ) that can enable action in the absence of a comprehensive political solution of scientific understanding.

Transparency and visibility of political decision-making are crucial, as scientific information alone will never rule out all uncertainties. In the introduction of this article, we classified GM crops as an unstructured problem characterised by disagreement about both facts and values, and about both the means and the ends of the technology. As probable solutions to this type of problem, a learning strategy and public debate were proposed (Table 1). We conclude that (formal) political debate should be added to this list. Based on the findings of this article, we endorse the claim of Van Oudheusden (2014), who emphasised the (leading) role of politics in responsible innovation. Finally, we note once more that the dynamics and characteristics of the debate on the safety of GMOs are not unique. Similar debates arise around other controversial technologies or developments that affect social values (e.g. shale gas, nuclear energy or vaccination programmes). The pointers and pitfalls identified in this article can therefore also be useful in the (political) debate about other controversial technologies.

Acknowledgements

This article is based on a report of the Netherlands Commission on Genetic Modification (COGEM). The opinions in this article do not necessarily reflect the opinion of COGEM. We would like to acknowledge all the experts of COGEM who were involved in the realisation of this report. Additionally, we are grateful to Annick de Vries (Rathenau Institute) and Lonneke Poort (Faculty of Law, University of Amsterdam) for their comments on draft versions of this article. Finally, we thank the participants of the workshop on food bioethics and food safety governance (2014) at China Agricultural University for their fruitful comments and questions.

Notes on Contributors

Ruth Mampuys is coordinator of the subcommittee on Ethics and Societal Aspects at the Netherlands Commission on Genetic Modification (COGEM).

Frans W.A. Brom is Head of Technology Assessment (Rathenau Instituut) and Professor Ethics of Technology Assessment (Utrecht University). Technology assessment and trust in pluralistic democratic societies are his research interests.

Notes

1. Regulation (EC) 1830/2003 concerning the traceability and labelling of genetically modified organisms and the traceability of food and feed products produced from genetically modified organisms and amending Directive 2001/18/EC.

2. When referring to politics in this paper, we address both the socio-political debate in the public realm on the impact of GM-technology as well as the regulatory-political debate related to the governance of GM-technology. Based on the description of Jeremy Waldron (‘that the felt need among members of a certain group for a common framework or decision or course of action on some matter, even in the face of disagreement about what that framework, or decision or course of action should be, are the circumstances of politics’ (Waldron 1999) and Chantal Mouffe (‘to envisage politics as a form of ‘agonistic pluralism’ in order to stress that in modern democratic politics, the crucial problem is how to transform antagonism into agnonism’ (Mouffe 2000), we are of the opinion that the ‘regulatory-political debate’ cannot be untangled from the ‘public socio-political debate’ that takes place outside of the regulatory framework. Wherever we refer to one of those in particular, this will be made explicit in the text.

3. A systematic overview of the arguments used in the debate can be found in a report of the Netherlands Commission on Genetic Modification: COGEM report CGM/131031-01.

4. www.grace-fp7.eu/en/home.

5. http://www.grace-fp7.eu/en/content/open-scientific-forum-public-debate-90-day-feeding-study-mon810-maize.

6. Directive (EU) 2015/412 of the European Parliament and of the Council of 11 March 2015 amending Directive 2001/18/EC as regards the possibility for the Member States to restrict or prohibit the cultivation of GMOs in their territory.