Nature Medicine | Spoonful of Medicine

Exit ghost

The article and accompanying editorial published in JAMA earlier this week regarding the issue of ghostwriting made, yet again, a strong point about the influence of pharmaceutical companies on the communication of science.

In a nutshell, a series of primary and review articles authored by Merck regarding their drug rofecoxib were allegedly written by communication firms that provide writing services for companies and individuals. In addition, some articles were signed by academics who allegedly did not have much to do with the study, but were invited by the company to lend their name (and their prestige) to the papers.

There seems to be something seriously wrong with this picture, but let’s go step by step. Is there anything wrong with companies that write papers for hire? The JAMA editorial seems to think so:

…there was no disclosure that the manuscript had been written by Scientific Therapeutics Information Inc, a company specializing in […] writing papers for a price. Perhaps some […] would see little or no harm in this failed disclosure because all other disclosures were made. However, if there was nothing to hide, why were the names (and affiliations) of the individuals who actually wrote at least the first draft of the manuscript omitted? Experienced authors know that the initial draft (in this case paid for by Merck) sets the tone for the manuscript. Moreover, it is unfair to the authors of the first draft not to provide them with credit for their work. Another problem with failing to disclose “ghost writers” is that there is a reasonable assumption that the principal investigator was involved with writing the manuscript from the beginning."


Why were the names omitted and credit was not given to the ghost writters? I would probably ask the opposite: Why should they be included? Scientific Therapeutics Information entered into an agreement with Merck according to which STI would write the paper and Merck would pay — that’s their compensation. Why should they be also authors? It’s like saying that papers should always include the names of all the technicians, as they did experiments that are reported in the manuscript. I’m sure many researchers would find this idea amusing.

Now, does the first draft really set the tone of the manuscript? What does that mean, anyway? In my experience, if the first draft of an article is bad, it will be discarded and people will start over. There’s no such thing as an article’s tone, at least not as something that the paper is born with from the time you first sit down in front of your computer and start typing.

And is there really an assumption that the principal investigator was involved with writing the manuscript from the beginning? I’d actually say the opposite — my assumption tends to be that the PI more often than not sees a draft written by the postdoc. In fact, many senior PIs would not dream of writing a first draft, but would impose it on junior researchers to get the paper into very good shape before they get involved.

This point is relevant to the second crime of which the Merck people stand accused — getting senior PIs to be authors in papers for which they didn’t do much work. These researchers are a little like those PIs who get involved in reading the paper once the draft is essentially final. And yet, we don’t often hear people asking for those PIs not to sign those papers.

Of course, this example is dumb — those PIs did contribute with the money to support the lab, and they may even have provided some intellectual input to the work (although, again, there are some senior PIs that let the lab in the hands of their postdocs, who pretty much decide by themselves what project they want to pursue). But even if the example is dumb, it does raise the point of what threshold must you cross to become the author of a paper.

Is it enough to be the owner of the lab for you to sign everything that comes out of it? Some would probably agree, even though no-one would dream of splitting the credit for painting the Sistine Chapel between Michelangelo and the Pope just because the latter “owned” the chapel. What if you “just” gave the mouse used for the experiments? Would you want to be author or not? Opinions will vary, but it’s not necessarily outrageous to appear in a paper for giving people such a crucial reagent. And what if you only gave intellectual input? Again, opinions will vary, but some people will value this kind of contribution more than giving the mouse or being the owner of the lab space.

So, why are we horrified about the idea that someone approached by a company to author a paper agrees to do it? After all, this kind of guest authors are often told that they are welcome to change anything they don’t like in the paper they are given. So, depending on how serious you are about this task, you may end up being really comfortable to give your name to a manuscript if you really worked hard on it. In fact, who hasn’t heard of cases in which someone ends up in someone else’s paper because of extensive discussions throughout the preparation of the manuscript?


I think that a crucial point in this discussion is included in the JAMA editorial: “…it is clear that at least some of the authors played little direct roles in the study or review, yet still allowed themselves to be named as authors”. Sure, but how little is “little”? In fact, some of the people who have been identified as guest authors have disputed the allegation precisely by stating that they did have a lot to do with the preparation of the manuscript. The same question goes for them: how much is “a lot”?

So, to my mind, the problem here are not so much the pharmaceutical companies for approaching people to author their papers, but how much people think they need to work to feel comfortable as authors of manuscripts of which they were not the leaders. Unfortunately, as I illustrated above, there’s no simple answer to this question, because there are many pathways that you can follow to end up being guest author in a paper — sheer seniority and ownership of the lab, sharing a key reagent or giving vaulable discussion.

The one thing that is different in this case is that money is exchanging hands — the pharma company allegedly offered money in exchange for the scientist’s name. So, if you just took the cash and signed the manuscript without even looking at the paper, sure, that’s clearly unethical. But if you did work on it, is it still inappropriate to charge a fee for your input? It may be against the rules of your institution, but I wonder if it’s inappropriate from a strictly ethical standpoint.

So, in the Merck case, if I were the Dean of any of the people who have been named as ghost authors, I would take the problem very seriously and launch an investigation to ascertain how much they really worked on the study. If they did, in fact, contributed enough (acknowledging that the definition of “enough” will vary depending on who you ask) and they didn’t break any rules of their institution, the institution needs to make this very clear to the community. And if it turns out that the authors really just took the money and ran, then disciplinary action must surely follow.

I share the view expressed in the JAMA editorial that integrity in medical research is paramount. However, I disagree with overarching, one-size-fits-all decisions about how to handle situations as complex as the issue of authorship. To my mind, these decisions can trivialize the problem by offering solutions that are a bit too simplistic and may end up being regarded as empty rhetoric.


  1. Report this comment

    Sergio Stagnaro MD said:

    Such as argument is since ever debatable. We must agree with the fact that no pharmaceutical company pays papers whose data state that their drugs either are useless or, worse, dangerous! In my opinion, two points are unavoidable in science writing in a honest way. Firstly, reader have to be in the condition to corroborate or falsify the referred data. Secondly, reader in doing that, must utilize generally CLINICAL tools, reliable in this essential task. I suggest my colleagues to go on a paramount way, which proved to be really relible during half of century. For instance: I was been told that new NADS drug is wonderful, with no side effects. Well! I bedside assess a lot of parameter values regarding joints inflammation, beside a numerous other, informing about antibody synthesis, CRP level, Reticulo-Endothelial Hyperfunction Syndrome, a.s.o. As far as, heart or kidney dangerous side effects are concerned, doctors can nowadays bedside evaluate biological system dysfunctions. In conclusion, authors may write what they wont, but after that, I shall controll all they results exactly.

  2. Report this comment

    El Dirque said:

    Why is this not criminally prosecuted?

  3. Report this comment

    Bo Yu said:

    To sign their names or not to sign their names, is the question. To sign their names results in lending prestige to the company. Not to sign their names leads to no advantage they can take of the rusulting paper issued by a “famous” journal which might be supported by the company. Most of PIs have to think about how long their prestige continues and how much “fund” their prestige could bring. Actually, this is a question of percentage and balance of truth.

  4. Report this comment

    JCL said:

    Thanks for your comments.

    I agree with Bo Yu that a scientist can, in principle, be guest author in papers only so many times before his signature is meaningless (if this is what Bo was trying to say…)

    It’s like a high-profile journal routinely publishing bad papers. It’s high-profile status would cease to be sooner rather than later.

    Regarding El Dirque’s comment, who would you prosecute and what’s the crime exactly? Hopefully you can elaborate on your point.

  5. Report this comment

    Chris Muller said:

    I believe very firmly that authorship is for those who can and will defend the paper before and after publication. It doesn’t matter if you gave a plasmid or a mouse, if you had long and interesting discussions in the restroom or a restaurant or if you paid for the building or a crucial instrument without which the research just could not have been conducted. Unfortunately, this is not how it works.

    Should a company be hired to write papers for a pharma company? Why not? The ultimate responsibility to verify all the facts remains on the shoulders of the pharma company. There is more honesty in a straight commercial transaction where services are rendered in return for a fee. But it is difficult to overlook the greed and dishonesty of any PI, and more so, a senior PI, who agrees to be an author for prestige issues if (s)he is not going to defend the work as his/her own. Who do they think they are that they are signing their names on these articles they have nothing to do with—Elvis Presley? This ain’t no autograph, people, we are talking about intellectual and technical work which is what brings respect to this profession.