17 Conflicts of Interest
I have always assumed that the subject of declaring a conflict of interest was straightforward and did not represent any significant explanation for those in the Biological Sciences. As it turns out, I was not only wrong, but I was very wrong. Like so many other aspects of science where policy is formed through evidence, there is a lot of potential for conflicts of interest in the Biological Sciences, and there are examples of where this has gone so wrong that even societies and their journals have been captured by large companies seeking to have science endorse their products at vast cost.
A conflict of interest (sometimes referred to as a competing interest) can be simply defined as any considerations that may compromise unbiased reporting in publication.
For scientists, this essentially means that their published material no longer adheres to the four tents of the scientific method: rigour, independence, transparency and reproducibility.
Typically, we think of Conflicts of Interest occurring when authors have a bias toward a particular viewpoint because, for example, it reflects more positively on the third parties from whom they received funding. But Conflicts of Interest might not only affect authors. In the scientific publication process, it also means that the journal editors (and potentially other gatekeepers) and reviewers may also hold conflicts of interest that could bias the way in which they handle peer review and decision making. These might be reviewers or editors aiding authors who have aligned interests. There can then be a myriad of reasons why competition between laboratories, old grudges or even petty prejudices may produce this bias (where a conscious bias represents a conflict of interest), but these aspects of reviewer bias are covered elsewhere (see Part IV). Instead, here I will focus on the ways in which corporate finances can impact publishing in the Biological Sciences, and perhaps the best way of doing this is through well documented examples. By bringing these to light my intention is to highlight the numerous ways in which Conflicts of Interest have already been found to involve the systematic manipulation of the publication process.
Table 17.1 contains some examples of Conflicts of Interest for different parties: authors, reviewers and editors. I have provided theoretical scenarios to show how alignment of these biases lead to manipulation of the publication process. In certain scenarios, these manipulations can be systematic and effectively lead to journal capture by interested parties. Note that in Table 17.1 it is only necessary for the Authors and Editors to show aligned bias in order to result in the systematic manipulation of the publication process. This only requires aligned reviewers if open reviews are obligatory in the journal. Editor oversight of reviewer choice is transparent with an open review system, and provides another reason for the move to Open Science.
|Reports bias in favour of a product which was the subject of the study because they received support from industry/manufacturer||Review is biased in favour of product as they receive similar funding||Chooses reviewers that are sympathetic to authors, and ignores reviews that criticise bias||When an editor is under the same bias as authors, the journal is effectively captured|
|Author has a personal or professional relationship with gatekeeper||Biased reviewers could be chosen by the editor||The editor has a personal or professional relationship with the Author and chooses reviewers that will be favourable and/or ignores criticism||Reciprocal agreements have been found where editors facilitate the publication of each others papers|
|The Author is a member of a lobbying of advocacy group||Biased reviewers could be chosen by the Editor||The Editor is sympathetic to the Authors cause and chooses reviewers that will be favourable and/or ignores criticism||The lobbying or advocacy group continues to publish in the journal where they have known sympathisers and advances their cause through scientific publications|
|Author is unbiased||Reviewer holds a grudge or rivalry with the Author and submits a biased review||Editor is unbiased||If the Conflict of Interest is undeclared by the reviewer, the editor could unwittingly collude in the grudge/rivalry|
|Author endorses a product in a paper for which they receive benefits from the manufacturer or supplier||Reviewer is unaware of the Conflict of Interest||Editor is unaware of the Conflict of Interest||The manufacturer gains endorsement for their product and the author benefits|
17.1 What does the Conflict of Interest statement mean?
To me, interpreting a Conflict of Interest statement is the nub of the problem with the current system. At present, there are a wide variety of ways in which Conflicts of Interest are stated. Many journals have them stated in the Acknowledgements, but without any declaration that the statement refers to a Conflict of Interest (see example below). Others have a statement entitled Conflict of Interest, but the declaration there is not elaborated. Instead the journal gatekeepers are saying ‘trust us’ to decide on your behalf. Simply by declaring that the author is a member of an organisation, or has an interest, funding or shareholding, doesn’t really explain the nature of the Conflict of Interest; i.e. the declaration is not a dispensation. More information is needed about exactly which part of the manuscript the Conflict of Interest relates to and why. For example, a statement on funding of research should include whether or not any oversight came from the funders, preferably with a link to more information on exactly how this did not or may have influenced the study.
In the case that an editor is an author, there should be a link to how the journal handles manuscripts when editors are authors.
Reviewers also need to declare whether or not they have a Conflict of Interest, and these should be visible along with their reviews (see more on the need for Open Reviews).
If there is a Conflict of Interest, then the category that this falls into should have a link to an explanation of journal policy, and the result of how policy was followed for this example and why the decision to publish was made.
Editors and board members (all gatekeepers) should have Conflict of Interest statements that are regularly updated on the journal website (together with the date when the last update was made). In the case where gatekeepers are industry employed or sponsored, we need a statement that explains exactly how the publication handles their involvement. For example, they may be excluded from any involvement in publications that include products from the same or rival companies. Ultimately, I can imagine how such people may be removed from much that happens in a publication, but that does not mean that I want such people excluded. Their role and participation in the scientific process is necessary, valued and important. Industry employees are likely to have already reviewed and discussed all possible Conflicts of Interest, and are well aware why their involvement necessitates scrutiny to maintain scientific integrity. Individuals without industry addresses, but with funding (past, present or only potential in the future) present a much greater challenge to journals.
For all concerned, we need to think about how to link income from funding to individuals without additional form filling. This can be simply achieved through independent mechanisms, such as ORCID, that keeps a user maintained list of research funding sources. Ideally, this would also have some oversight from the researchers’ institution.
17.1.1 No Conflict of Interest
By declaring that they have no Conflict of Interest, authors are stating that their study is independent of any potential for bias.
The declaration is very important as authors need to think carefully about how possibly it could be conceived that a Conflict of Interest has occurred. If a declaration of no Conflict of Interest is made and retrospectively the journal or editor suspects that this is not correct, COPE has documents of how this should be handled by their members after submission of a manuscript (COPE, 2019a) and after publication of a paper (COPE, 2019b).
17.2 Examples of Conflict of Interest
Conflicts of Interest are perhaps most obviously manifest when industry sponsors research on their own products occurs. But industry is not the only source for Conflict of Interests in science. Governments determine policy in part through advice and evidence received from scientists, and there are examples where individual scientists will provide advice that favours their own standing or that of their laboratory or institution (Lackey, 2007).
17.3 Atrazine, Syngenta and their sponsored research
The story of how a large chemical company financed studies and only allowed certain results to be published has been well reported with documents released in court that shows how they took this further to try to discredit those who had the potential to expose them. The case is so clear cut, that anyone reading the facts might think that the chemical in question would have been widely banned from use globally and the company sued into oblivion. However, that’s not how the world works, and despite overwhelming evidence, the company continues to manufacture and sell this chemical which continues to make large profits for its shareholders. You can read some comprehensive accounts of the story by Rohr and McCoy (2018) and Hayes (2004).
Atrazine is one of the most widely used herbicides in the world. Atrazine is made by Syngenta AG, a Swiss based agrochemical company and subsidiary of National Chemical Corporation (ChemChina) who acquired it in 2017. There are many studies (reviewed by Hayes, 2004; Rohr & McCoy, 2010) that demonstrate harmful effects of Atrazine in freshwater vertebrates which (among other things) can increase mortality, affect behaviour and physiology, suppress immunity and increases infections, disrupt the endocrine system, cause gonadal abnormalities and induces community-wide, indirect effects (see Rohr & McCoy, 2010). Perhaps most famously, Atrazine has been found to demasculinize frogs at concentrations lower than those considered safe for human drinking water (Hayes et al., 2010).
17.3.1 EcoRisk and their Conflicts of Interest
Syngenta commissioned a lot of work on Atrazine, as we might expect. They did this largely through a third party company “EcoRisk Inc.” that maintained links with the scientists sponsored so that EcoRisk had oversight into the experimental design, and a say on which results could get published. This means that effectively Syngenta could control whether or not results that might show Atrazine in a bad light were submitted for publication or not.
The scientists that were paid received large sums of money which, because of the way that universities are administered, meant that they became prominent in their institutions and communities. This includes becoming integral parts of the gatekeeping societies for the journals that were supposed to be carrying unbiased results following peer review. This in turn meant that editors could influence the results of peer review such that papers that contained major flaws were still published.
This systematic manipulation of the publication process led to the existence of contrary reports in the scientific literature, and meant that Syngenta could continue to argue that results were not conclusive (see Michaels & Monforton, 2005) and that the herbicide should continue to be licenced (which it is in the USA and other parts of the world). When several of these Syngenta paid scientists sat as gatekeepers of the journal Environmental Toxicology and Chemistry it meant that it became a safe place for submissions on Atrazine and the journal (and arguably the Society of of Environmental Toxicology and Chemistry: SETAC) had effectively been captured. Reading the affiliations of SETAC board members is like reading a directory of the world’s largest chemical companies. This is not necessarily a bad thing, and I should point out that SETAC does have a statement on Conflicts of Interest, but none of the declarations made appear to be open (i.e. accessible).
The journals in which these articles are published call for statements on Conflicts of Interest in both the letter to the editor and in the acknowledgements. However, when they are printed, there is no clear statement titled “Conflicts of Interest”. For example, Du Preez et al. (2008) state in their acknowledgements:
17.3.2 Example acknowledgement
“This research was conducted under the oversight of the Atrazine Endocrine Ecological Risk Assessment Panel, Ecorisk, Inc., Ferndale, WA with a grant from Syngenta Crop Protection, Inc.” (Du Preez et al., 2008)
Given that the work was entitled: “Reproduction, larval growth, and reproductive development in African clawed frogs (Xenopus laevis) exposed to atrazine”, we should expect that anyone reading an industry sponsored study on the effects of their own chemical should have it drawn to their attention that there is a very definite Conflict of Interest occurring between the authors, who were paid by industry and the results of their study on the risks of the industry produced herbicide. Add to this the results which state “no effects of Atrazine” over and over in the abstract, and alarm bells should be ringing for anyone involved in gatekeeping the journal Chemosphere. While this work might have had the highest level of integrity, any casual reader that came across it would have no knowledge of any extra care that had been taken in the experimental design, avoidance of confirmation bias in the analyses and the steps taken to avoid clear Conflicts of Interest in the editorial and review process. Indeed, in my opinion, the fact that all of this information is missing should mean that any reader will treat the publication with suspicion.
Despite the presence of a large body of inconclusive work on Atrazine, an independent review covering all publications still conclusively demonstrates how harmful it is. When Rorh and McCoy (2010) reviewed the literature objectively, it did not matter whether the industry sponsored studies were included, the negative effects of Atrazine still outweighed the inconclusive studies (Michaels & Monforton, 2005). Interestingly, when the industry sponsored team reviewed the same publications, their conclusion (published in Critical Reviews in Toxicology) was completely different (Solomon et al., 2008). Note that this is not a case of scientists disagreeing (which is what Syngenta might want you to think), this is a case of journals failing to ensure that integrity in science outweighs the funding of those with a Conflict of Interest.
17.4 How to interpret a Conflict of Interest statement
Our immediate problem is knowing how to interpret statements that suggest or imply a Conflict of Interest.
Open peer review and the clearly stated Conflicts of Interest of all those involved (authors, reviewers and gatekeepers) needs to be prominent in any industry sponsored studies. Over and above this, we should see a system where money from industry is never tied to any oversight of the results that scientists produce. There are systems that could be used to effectively separate funders and the independence of the results that funded scientists produce, but for as long as it remains in the interest of gatekeepers, we are likely to see the current system perpetuate into the future. While I would normally turn to COPE for a clear lead on Conflicts of Interest, their overview of systematic manipulation of the publication process is more concerned with other issues (such as ghost authorship, papermills and plagiarism, all of which are tackled elsewhere in this book) than capture of journals and societies by industry. Moreover, they give no clear instructions on how to declare a Conflict of Interest within the publication. If it’s not prominent to the reader, then the publisher is effectively hiding it (like in the example above).
This is not really a question of balance: we do not need an industry view and an independent view. At its heart, science must adhere to the four tenets of the scientific method: rigour, independence, transparency and reproducibility. In the case of industry sponsored research, we need both independence and transparency to be demonstrable by the authors of any research. Industry cannot have a role to oversee research on their or other products, and in my opinion, the declaration of an industry led oversight (as in the example of Du Preez et al., 2008 above), while somewhat transparent, doesn’t actually inform us how to interpret that information. To me, this appears to be a declaration of an absence of independence, and as this would violate one of the four central tenets of the scientific method - the editors should have insisted on a detailed declaration of exactly what this oversight consisted of.
17.4.1 Can we rely on gatekeepers to make decisions on our behalf?
I would say that we certainly need to trust gatekeepers to maintain the integrity of their own journals. They are best placed to make these judgements, but making such decisions transparent and open for all readers should be the standard that all editors aspire to.
Hence, in the example above, it may well be that gatekeepers at the journal Chemosphere felt happy that detailed Conflicts of Interest statements submitted by the authors to them were sufficient for them to believe that the research remained independent. Remember that many companies don’t rely on contradictory evidence being published, it is enough to publish inconclusive evidence (Michaels & Monforton, 2005).
The danger comes when the journal gatekeepers and authors have aligned conflicts (see Table 17.1). Then their judgement about the independence will likely be impaired. When we look back at publications accepted in the past, how easily can we tell whether or not the editor back then had an aligned conflict of interest? For me, it is easier to accept contemporary decisions that we can question if need be. But looking back on literature in the past leads to questions that can likely never be answered. My opinion is that we need this information now so that readers in the future are better able to evaluate editorial decisions made today. This will protect authors who are meticulously unbiased but also sponsored by industry from suspicion and scurrilous accusations, as much as it does the integrity of the journal and the scientific publishing process.
As with other aspects of Open Science, we need documentation that is available to everyone both at the time that the work is published and, perhaps more importantly, in the future.
17.4.2 Tackling journal capture from the past
We are left with a number of journals that have previously been captured, and while their current editorial team may have the greatest integrity, the past systematic manipulation of the publication process requires review to determine whether articles published under biased editorship need retraction.
If allegations around systematic manipulation of the publication process surface, then the scholarly society (or the publishers if no society is involved) needs to appoint a committee to review editorial decisions made to assess whether Conflicts of Interest arose. Readers need to be informed via an Expression of Concern (see Table 33.1) posted on the journal pages relating to each paper. If appropriate documentation is absent, then authors of the papers in question will need to engage with the committee in order to assess each case. If a Conflict of Interest is uncovered then a full retraction may be needed, together with a notification of retraction. COPE (2018c) have a useful process to guide societies and publishers on the systematic manipulation of the publication process. This will need to be adapted to the particular system in the case of an industry captured journal.
17.5 An Open Science approach can solve the crisis in Conflicts of Interest
I would argue that Conflicts of Interest need to be honest and openly declared for everyone, but that they should not exclude individuals or groups from having a voice in science. Clear and open policies should be drawn up for journals to follow so that authors, reviewers and editors all know how their Conflict of Interest will be handled. It is important that such policies are not punitive, or deter the declaration of a Conflict of Interest in any way.
While some authors have suggested a top down solution to determining the influence of industry sponsored research (see Rohr & McCoy, 2010), and others focus on an publisher based approach for issues around Conflict of Interest (COPE, 2019a). I prefer a science-centric approach that does not leave us vulnerable to the vagaries of national or international politics. Happily, Open Science has the tools to accommodate industry sponsored research while maintaining scientific independence and rigour.
We know that the movement towards Open Science includes the preregistration of experiments (Forstmeier, Wagenmakers & Parker, 2017; Nosek et al., 2018). Any industry that sincerely wanted to fund a scientific study (i.e. one that adheres to the four tenets of the scientific method), would require that the experiment be planned. Registration of the plans for an experiment can be submitted for independent peer review, and lodged in an open access and independent repository. This would increase the rigour of the work as it would also ensure that as it ensures against confirmation bias. Ideally, this level of rigour should not be reserved for industry sponsored research.
This approach does not need to stop with experimental science, but can extend to reviews and opinions where the objectives and methods are clearly predetermined and their veracity independently assessed.
Of course, we still rely on scientists to adhere to the four central tenets of science. We recognise that scientists are humans, and that there may well be reasons why they want to introduce bias into their work (Zvereva & Kozlov, 2021). Preregistration of projects and a review of these together with submitted work would substantially improve their work, and allow them to point to independence over their sponsors. Scientists avoid preregistration as it is extra work, but in the case of Conflicts of Interest, it seems clear that this is the only real way of removing the potential for Conflict of Interest corruption.