What should we do with scientific ghost writing?

Ghostwriting in peer-reviewed healthcare literature has been gaining quite a bit of attention. PLoS has even created a public web site to document this deplorable practice by one pharmaceutical manufacturer. But what is this really all about? Is it another tempest in a teapot or a dire threat to scientific integrity? The answer is probably a little bit of both.

Before going further, a couple of disclaimers. I have published a number of studies supported with research funding from manufacturers. In only two instances I had manuscript preparation assistance from a medical writer, and each was acknowledged in the paper. I generally do not like having someone else draft the paper, even a co-author, and prefer to do so myself (control issues, clearly). I also agree that having a paper written by a medical writer and slapping an academic's name on it willi-nilli is an indefensible practice. To be sure, it is indefensible on the parts of both the sponsor and the academic, and as such should be systematically routed out.  

Now that we have gotten that out of the way, let's focus on the topic at hand. Here is the definition of a ghostwriter from my favorite populist source, Wikipedia:

A ghostwriter is a professional writer who is paid to write books, articles, stories, reports, or other texts that are officially credited to another person. Celebrities, executives, and political leaders often hire ghostwriters to draft or edit autobiographies, magazine articles, or other written material. In music, ghostwriters are used in film score composition, as well as in pop music such as Top 40, country, and hip-hop. The ghostwriter is sometimes acknowledged by the author or publisher for his or her writing services. [emphasis mine]

Clearly, ghostwriting is a pervasive and acceptable practice in literature, music, politics and even journalism. So, why is there such a furor over it in medicine? Well, the answer is "bias". Here is an example cited by the PLoS editors of an illustrative correspondence between a sponsor and the medical writing company:

An email from a writer employed by the medical writing company, DesignWrite, to employees of Wyeth, the company that performed the study, and Parthenon (another medical writing company) on November 10, 2003 concerning manuscripts on Totelle (a brand of hormone replacement therapy manufactured by Wyeth) tells the story concisely. “Thanks to all who have reviewed and approved the manuscripts… I have received no word on authors for the Totelle 2 mg bone manuscript P3(2), and need input on this matter before this manuscript can move forwards.” [our emphasis added]

On its own, the quote is damning on the one hand, but totally explicable on the other: in large trials not all investigators are listed as authors on the paper, and this my be what the e-mail is querying. I do not think that this was the case here, however. As for generating review papers with authorship to be determined at a later date, the case is less clear. Should the manufacturer be able to be the originator of an idea on what information gaps need to be filled? Or should these always come from academics?

Consider this: manufacturer organizations are not monolithic. It is unfortunate that sales and marketing departments are usually oversized, powerful and most visible. Less visible are R&D groups, Clinical Affairs divisions responsible for Phases 1-3 work, and Medical Affairs departments, whose role is to shepherd the post-approval clinical development. It is the Medical Affairs, staffed by healthcare professionals including physicians, pharmacists and others, that primarily interacts with academics to develop ideas that presumably have both, academic integrity and strategic merit for the company. Smart companies understand the importance of having a firewall between the sales and marketing organizations and the Medical Affairs in order to safeguard against marketing abuses couched as science. Under these circumstances, even if a paper starts out within the walls of the company, there has been a heavy imprimatur of the academic collaborators, who ultimately, along with the company authors, take full responsibility for the content of the publication. My point is that not all papers "commissioned" by the industry need to be thrown out.

But what is the role of a medical writer in these endeavors? In my experience, medical writers develop an outline and the first draft of the manuscript based on the data at hand. The quality of this draft can be quite variable depending on the writer's experience. An ideal medical writer is not only knowledgeable about the particular field of investigation, but also knows how to write. Unfortunately, these two criteria are rarely met either singly or together. So, the draft that the authors get usually goes from incoherent to rambling and requires a great deal of intervention at both the scientific and the writing levels. So, if the implication is that the papers get accepted into high impact journals as written by the medical writers without much shaping and editing by the academics, there is something very wrong with these journals' review practices: it is easy to spot an unedited piece of hired writing a mile away! 
      
The World Association of Medical Editors defines ghostwriting thus:

Ghost authorship exists when someone has made substantial contributions to writing a manuscript and this role is not mentioned in the manuscript itself. WAME considers ghost authorship dishonest and unacceptable. Ghost authors generally work on behalf of companies, or agents acting for those companies, with a commercial interest in the topic, and this compounds the problem. For example, a writer employed by a commercial company may prepare an article, then invite an expert in the field to submit the work, perhaps with minor revisions, under his or her own name. The submitting author may be paid, directly or indirectly, for this service. In other circumstances, investigators may pay a professional writer to help them prepare their article but not mention this assistance, gaining credit for writing they have not done. Although editors seek to avoid publication of ghost written articles, these articles are often very difficult to detect.

Again, I have to say that ghost writing that I have been privy to as a reviewer is really not that hard to spot. But of course this view may be the result of misclassification on my part, as those papers that are well written do not elicit the same suspicion from me. I guess to be more accurate I need to state that at least some ghost written papers are easy to spot.

So, what is the sum total here? So long as there is a profit motive on the part of the manufacturers (sorry, but this is the way capitalism works today, right or wrong), they will have a hand in what gets funded. Short of scrapping the pharmaceutical business model as it exists today, we need to come up with transparent best practices that everyone can live with. Companies do and will commission papers. It is unwise for the sales and marketing departments to be the originators of these ideas because of the unmitigated conflict of interest. It is best for Medical Affairs professionals, insulated from sales and marketing, to interact with the academic experts to develop and execute studies and reporting of the data. Medical writers should be used as such only if they bring writing and content expertise to the table, and they need to be acknowledged either as a co-author (provided they meet the criteria) or as someone who has made a contribution to the drafting of the manuscript. If they are unable to contribute usefully, they should be used only for such administrative tasks as formatting the manuscript and references, collating comments and submitting the paper.

On their part, journals need to set higher standards for the reporting, disclosure and reviewers' comments. Editors and reviewers should be on alert for poorly written review papers with a marketing message -- from what I have seen, these are just not that subtle. Ultimately, the responsibility for the integrity of scientific discourse lies with all of the stakeholders. Lest we blow our scientific reputations with the public, we have got to fix what is wrong and come up with a solution incorporating those pieces that work and do not compromise our professional ethics. We do not need to throw the baby out with the bath water; a rational solution may be much closer than we think.