Davies, John, Cameron,
Douglas, Drucker, Ernest (2002), Damned and be Unpublished.
© Copyright 2002 John Davies. All rights reserved.
Damned and be Unpublished?
John Davies, Douglas Cameron, Ernest Drucker
The Farmington consensus and the " International Society
of Addiction Journal Editors".
This article is derived from two editorials which have been published in Addiction Research and Theory (AR&T) over the past eighteen months (Cameron, D. 2001; and Davies, J.B., Drucker, E. & Cameron, D. 2002). They elucidate the standpoint of that journal with regard to the issue of conflict of interest in scientific research. We have found ourselves needing to be explicit because of the presence of a document called "The Farmington Consensus" (1997) and because of the creation of a body called the "International Society of Addiction Journal Editors".
In the matter of publication of scientific findings, absolute transparency seems to be the current credo. Those of us who wish to publish our results and have conflicts of interest should declare them. If we receive funding from some organisation which is funded by a sect which believes that we are descended from people from outer space, we should make that clear. And having done so, our views will be judged accordingly. "Accordingly" means, of course, given less credence than if our source of funding was more reputable, was, say, a government funded research council or a worthy charity or public trust. We are told that conflict of interest is not in itself a crime against science. Non-declaration of such a conflict is. But who can judge which sources of funding are worthy and which are not, which conflicts of interest need to be declared and which do not? That question is particularly cogent in our field, drug use.
Recently, a new society has been formed; the International Society of Addiction Journal Editors (ISAJE). On the face of it, this is a laudable enterprise aimed at bringing together editors of the more than 40 journals that deal with the scientific and intellectual issues associated with drugs and the problems that we face in dealing with them. We, the Editors of Addiction Research and Theory, have declined to join this organisation, and these are the reasons we gave to our readers why we opted not to join.
In order to become members of that society, journal editors have to make a set of declarations; some of which are reasonable and benign, but one of which, in our view, is not. The declarations are as follows.
- (The journal) is a peer-reviewed journal, which submits most of the articles it publishes for review by experts outside its own staff;
- (The journal) supports the Farmington Consensus;
- (The journal) is not subject to any undeclared institutional conflict of interest.
We have no problem with points a) and c) but we have difficulties with b), the requirement that we endorse the Farmington Consensus (1997) an informal agreement arrived at by a number of addiction journal editors.
Of course if examined on face value, the Farmington Consensus seems worthy. Who could argue with a statement from a group which has as its declared aim to enhance the quality of our endeavours in this multidisciplinary field? Like others in public we are against sin. But in private, we have severe reservations about the relevance and legitimacy of the so-called evolving process and its attempt to "define the basis for shared identity, commitment and purpose among journals publishing in the field of psychoactive substance use and associated problems". [ibid] In particular, we worry about Clause 3 of the Commitment to the peer review process (1.3).
Referees should be asked to declare to the editor if they have a conflict of interest in relation to the material which they are invited to review, and if in doubt they should consult the editor. We define "'conflict of interest "as a situation in which professional, personal or financial considerations could be seen by a fair-minded person as potentially in conflict with independence of judgement. Conflict of interest is not in itself a wrongdoing. [ibid]
And about Clause 4 of the Expectations of Authors (2.4).
Conflicts of interest experienced by authors: Authors should declare to the editor if their relationship with any type of funding source might be fairly construed as exposing them to potential conflict of interest. [ibid]
First, there appears to us to be an unspoken implication behind the Consensus that certain researchers produce research which is flawed, biased or downright fraudulent as a consequence of their being dependent on or beholden to certain sources of funding, Big Booze (BB) and Big Tobacco (BT) being the two prime sources of concern. Increasingly, of course, the pharmaceutical industry would also be a candidate for potential conflicts, as we become aware that some of their products can be a major source of problems. It goes without saying that some of the manoeuvres perpetrated by BB and BT do not bear close scrutiny, and that a degree of scepticism or even censure is entirely appropriate with respect to many of their research and marketing activities. It does not follow, however, that a similar bias is absent from research funded by "nobler" bodies such as the cancer foundations, health institutes, research councils and so on. The assumption that these latter studies are carried out by "scientists" without bias or prejudice, whose aim is merely to find out the "truth" that emerges inevitably from data ("the disinterested and objective search for truth") is a dismal and sorry belief in this day and age; and yet this is exactly the position advocated by the prime movers of the new society.
So, the purpose of the Farmington Consensus is, bluntly, to separate the good guys from the bad guys. However, contemporary philosophies of science (as distinct from 17th century views) see knowledge as relative, and researchers as motivated individuals whose actions are coloured by their own interests and beliefs. Obviously, the fact that research is always and inevitably carried out by individuals with their own agendas and biases, and is influenced fundamentally by context, creates problems for what we loosely refer to as "science", but these problems have to be dealt with rather than simply dismissed. It is clearly utter nonsense to suggest that research takes place in no context and from no point of view at all.
In the health literature it is possible to find reference to the hired lackeys of Big Tobacco (or some such phrase); sentiments which some of us would probably share in specific cases. However, one cannot dismiss all research from such sources simply because some of it fails to stand up to scrutiny; just as we cannot accept all research from more respected sources, simply because of its provenance. If the good guys are characterised as disinterested and objective, there are no good guys unless they are zombies. To be alive is to have views and opinions, which guide what one says and does.
The second issue of concern is much simpler. Recent events show clearly just how the Farmington Consensus can be used, and the purposes to which it can be put by the members of the ISAJE; that is, as a tool with which to attack individuals who hold new or unfashionable ideas, and in the process to endanger careers and destroy reputations. An addiction journal, the editor of which has been a prime mover both in Farmington and the proposed new Society, recently took the remarkable step of 'unpublishing' a previously published article (Elleman-Jensen, P. 1991) on the basis that an undeclared conflict of interest had been discovered ("the paper may have had an undeclared tobacco industry connection") which 'therefore' disqualified the article in the interests of "the integrity of science". No statements were made about the substance, accuracy or value of the study - merely that its author was somehow tainted. This attempt at post hoc censorship strikes us as both unwise, and as having ramifications which are bizarre. Once something is published in the public domain, it stays in the public domain and can get cited by others in their papers. Should these be "unpublished" too? Should we expect to see in the list of publications in our Curricula Vitae entries such as "Published 1988, Unpublished 2002"? It is clear to us that a paper cannot be removed from the public domain by the simple expedient of stating that it never, or should never, have appeared. One wonders just where such a process, once started, might stop; and how far the criteria for this might extend.
In the event, since the author of the now "unquotable" study has died and is not able to defend himself, ex-colleagues who are both physically and academically alive leapt to his defence, and pointed out that the information about a supposed conflict of interest on which this decision was based was in fact incorrect (Pedersen, K.M & Christiansen, T. 2002). Had these ex-colleagues not come to the rescue, a disturbing post mortem destruction of a reputation would have taken place. Following that defence there was an attempt to justify the 'unpublishing' course of action embedded in the letters page of a later issue of that journal (Edwards, G. 2002). Therefore, presumably, the paper is still meant to be unpublished and uncitable.
More recently, in the same journal, a second related instance has come to light concerning a book which gives a critical appraisal of the notion of nicotine addiction. (Frenk, H. & Dar, R. 2000). The text includes methodologically and historically important criticism of some of the "landmark" papers, including instances of what appears to be, to all intents and purposes, fraudulent practise (failing to report data that do not support the argument, deleting data from subjects that do not behave in the way "required" etc). The book grew out of interest sparked in the authors who had done a partial review of the literature which was commissioned by a firm of lawyers one of whose clients was a tobacco company. That connection was used by the Editors of Addiction as the justification for a major editorial outcry (N.B. not a book review) under the banner "Another mirror shattered? Tobacco industry involvement suspected in a book which claims that nicotine is not addictive" (Edwards, G., Babor, T., Hall, W. West, R. 2002). In contrast to that response, on receipt of the book at AR & T, we delved into its content, reviewed it, and find it to be an act of considerable scholarship, well worth the reading . We also note on page 3 of the book a specific authors' statement "that in arguing that nicotine is not addictive, we have no intention of praising the merits of smoking;" and on page 4, "We believe it would be rational for everyone to avoid or minimize smoking because of the short-term and long-term effects on health." Nonetheless, the authors found themselves on the receiving end of what can only be called a self-righteous crusade, damning their book because some of the initial source material was commissioned by a firm of solicitors with tobacco connections. We are invited to reject the content in its entirety on that basis. In our view this amounts to nothing less than an attempt to suppress a coherently argued position and thus to suppress an ensuing open debate which in our view is the very stuff of scientific discourse.
Readers should also note that the book was not a submitted paper. Whilst journal editors can of course comment on any book they wish, normally one would hope through a proper book-review process, the editors of Addiction might well consider the fact that Farmington applies specifically to publications submitted to addiction journals and not to the world of publishing as a whole.
We now see that Farmingtoncan be used to disqualify entire bodies of argument and criticism with which certain people do not agree, on grounds that have nothing to do with the scientific merit of the work; and to launch retrospective witch-hunts whereby previously published work can also be disqualified. In our view, this is morally indefensible. And where would this end? The Frenk and Dar book was not funded by tobacco (the authors are academics supported by their research and teaching). The solicitors who paid for the original literature review had tobacco connections, but the authors simply provided expert evidence as academics often do, and found the literature wanting. Suppose the solicitors in question had no tobacco connections, but were in partnership with a firm that did? Or what if one of their directors worked for two firms, one of which had such connections? In our minds at least, we can almost hear the sighs of relief from the "good guys" that a connection had been established. Any connection will do, no matter how remote, just so long as this body of evidence can be disqualified at source without having to comment on the content.
In the US many researchers have experienced considerable difficulty getting funding for studies that employ Harm Reduction - a term that despite being official policy in many countries, is still proscribed by the US federal government and its leading drug research agency NIDA - which funds 80% of the world's research in illicit drugs. This applies also to "recreational drug use" (which could well constitute 80% of all illicit drug use); studies of "controlled" alcohol use among former alcoholics (which offends 12 steppers); and harm reduction approaches to "controlled" (i.e. low level) tobacco smoking, which was termed "a dangerous idea" in the peer review comments to a grant application made by one of us (ED) to the US National Cancer Institute. In addition to restrictions of funding based on the clash between evidence and official ideology, we also face the difficulties of doing justice to the actual lives of the majority of drug users - especially the positive experiences and effects that motivate much non- problematic drug use (e.g. medical marijuana). Will personal histories of drug use soon become a disqualifying characteristic - about which assurances, disclaimers, and (soon) urine samples will be expected along with the submission of an article? And what about the behaviour of one's associates? This is not an idle concern in the US , where a recent Supreme Court ruling upholds the eviction from public housing of anyone whose family member (even if living elsewhere) was found to be using illicit drugs. With the newly launced Anti Drug campaign linking illicit use to support of terrorism , the stage is set for all sorts of drug related witch-hunts - now in the name of national security.
The third and final point is just common sense. We at AR and T wish to maintain our editorial independence without decisions about what we may or may not publish being determined by others. We are uneasy about monolithic across-the-board agreements about what can and cannot be published and believe these have a poor history (see for instance John Milton's "Areopagitica" speech of 1644). At a practical level we have to accept that bias, distortion of facts, good research, bad research, can come from any source. If authors wish to declare conflicts of interest we publish them readily, but we do not insist on it. Everyone has a conflict of interest somewhere, and Farmington merely encourages the search for it by others with their own vested interests or whenever something is published with which they disagree. Farmington is thus a controlling and stifling influence; and we fear it is just those ideas which do not fit into the easy "addiction" consensus which will be stifled. Science is always a humbling experience, and the still fairly primitive state of our understanding and effective control of the nastiest features of drug use and addiction suggest we have a long way to go before we have mastered that domain. New and unorthodox ideas that suggest new directions, or inspire us to look again at received wisdoms, are desperately needed. Meanwhile, in our journal we have decided that we will continue to use the peer review process to screen out poor quality research but we would prefer our readers to exercise their own judgement about the meaning and significance of work published in Addiction Research and Theory, rather than being denied the opportunity to do so as a consequence of our Editorial Board handing over its sovereign rights to someone else.
- In 1644 the poet John Milton published the text of his Areopagitica speech, an appeal to Parliament to rescind a Press Licensing Order. This Order sought to bring publishing under governmental control, by creating a body of censors to whom all work had to be submitted for approval. Milton argued elegantly that the order represented attempts by the state to control thought
Cameron, D. (2001) Editorial: The End of the Peer Show? Addiction Research and Theory 9:187-192.
Davies, J.B., Drucker, E. & Cameron, D. (2002) Editorial: The Farmington Consensus: Guilt by Association. Addiction Research and Theory 10: 329-334.
Edwards, G., Babor, T., Hall, W. & West, R. (2002) Editorial: Another mirror shattered? Tobacco industry involvement suspected in a book which claims that nicotine is not addictive. Addiction 97: 1-5.
Edwards, G. (2002) Addiction's decision to withdraw a published paper from citation on the grounds of an undisclosed conflict of interest. Addiction 97: 756 - 758.
Ellemann-Jensen, P. (1991) The social costs of smoking revisited. British Journal of Addiction 86: 957-966.
The Farmington Consensus (1997) Addiction 92: 1617-1618.
Frenk. H and Dar. R (2000) A Critique of Nicotine Addiction. The Netherlands, Kluwer.
Pedersen, K.M & Christiansen, T. (2002) A fair hearing or academic Kangaroo court. Addiction 97: 227- 229.