A ghostly undertaking
Many years ago I was the project manager of a large clinical study. I ran the study, did most of the data management and a good deal of the analysis, and then left the company for pastures new. Someone else came on board and wrote up the results for publication. My name was added to the list of authors, but I was never shown any drafts of the manuscript, and indeed didn’t see the paper in any form until it was published. Over the intervening years my role has been reversed. I have written up a lot of studies for publication, and been paid to do it. But in most cases I didn’t have anything to do with the conduct of the study.
So does either of these examples put me beyond the pale in clinical science? Lots of people will say yes, they both do. So-called ‘ghost writing’ is currently receiving a lot of criticism, and I am not convinced that all of it is justified. But before I try to defend the practice, let’s look at some facts.
In my first example, I knew the data inside out, the findings were very clear, and I didn’t have any problem with how they were presented in the paper. In fact the results were quite surprising. I don’t think anybody would say that I should not have been an author – indeed I drafted a lot of the methods section before the other writer took over. Neither should the new writer have been denied the credit, as they carried out the rest of the statistics. So I don’t think either of us can be called a ghost writer.
“So-called ‘ghost writing’ is currently receiving a lot of criticism, and I am not convinced that all of it is justified.”
But I do think there was a significant shortcoming on the matter of review. Interpretation is particularly vulnerable to bias in scientific publications. As I say, in this case the results were very clear, although they did overturn some prior assumptions. Looking back, I should have had the opportunity to challenge the interpretation. Indeed, I should have insisted on it, mea culpa.
So what about the much larger number of papers that I have ghost written?
I have never had any ethical worries with doing this, and I’ll explain why. Let me go back to the period of my first example. At that time I inherited a clutch of clinical trials set up by a predecessor. Bear in mind that there was no Good Clinical Practice in those days, or at least it had been launched by the FDA but my company had not heard about it. At least there was a written agreement with each investigator, and as most of the trials were single centre phase IV studies most sites were self contained.
The agreement in most cases stated that the investigator was responsible for writing up the results. Of course, nobody had thought about where they were going to get the resource for data management and statistics, and in several cases the investigators had difficulty in doing it. I well remember one study which never was written up, because there was no data analysis function available. By modern standards that was pretty reprehensible, as publication of results is a hot topic these days. There is no doubt that the sponsor should have provided the necessary resource, even if that involved engaging contract biometrics and medical writing staff. In the event the sponsor pulled the plug and the investigator didn’t get paid.
But I want to contrast that case with others with which I have been presented, in which there was a statistical report but no manuscript for publication yet written. The bald fact is that, very often, the only way the study is going to be published is to engage a specialist to do it.
“The bald fact is that, very often, the only way the study is going to be published is to engage a specialist to do it.”
There are several reasons for this. Clinicians have enormous pressures on their time, especially in the UK, and it is very hard work even to get them to participate actively in clinical research. When a study completes, there is a general sigh of relief and people naturally look to other things. Writing it up is the Cinderella of the whole process. At that stage, we all need to remember why we started the trial in the first place.
But not all investigators wash their hands of a completed study. In many cases there are several participants, all with an avid interest in the subject area, and coordinating all their perspectives on the results can be enormously onerous. Sometimes the job needs a view from outside, making it easier to apply objectivity. In addition, is it reasonable to assume that a top medical scientist is necessarily a convincing writer?
In the course of critically reviewing the literature, I find that the quality of expression can be very poor. I am surprised that this survives peer review, because sometimes it really is difficult to understand what is being said. This is especially true of abstracts of papers and of oral (or poster) presentations at conferences. The problem is that citation listing services, of which PubMed is the best known, obviously only publish the abstracts for space reasons, and if the abstract doesn’t accurately reflect the full text then understanding can be impaired.
Just the other day I received a paper on a rather esoteric subject. Bearing in mind that the authors claimed to be the first to apply a particular technique to their field, the methods section was astonishingly brief – only 201 words in a 3,500 word paper. At the same time the explanation of what they did was pretty much unintelligible (at least to me!). This appeared in a supposedly peer reviewed journal, in the stable of a major publisher. I can only assume that the reviewers were clairvoyant, or had some other way of gaining understanding from so few words.
I have seen many critical reports citing papers supposedly authored by key opinion leaders, who had nothing at all to do with either running the study or writing up the results. I am well aware that this happens in academic papers as well as in commercially sponsored ones, the professor always used to get their name on whatever paper was produced by their department.
“I have seen many critical reports citing papers supposedly authored by key opinion leaders, who had nothing at all to do with either running the study or writing up the results.”
Over some 20 years I have never been involved with any such paper, and have always liaised closely with all authors, which entails multiple iterations of the text until they are happy with it. On some occasions I have also been drawn into dialogue with the journal editors, putting me into the middle of a complex web of communication. I don’t for a moment think that any of this compromises the accuracy or quality of the finished paper. I really can’t remember anyone trying to force me to use words that didn’t reflect the data appropriately, or to bias the presentation. In other areas of work I have (rarely) had to warn clients that they are not following my professional advice, but in medical writing ethical standards seem to be quite good.
Of course, my experience need not be representative of standards in general, and there are clearly improvements that can be made. In particular I think that if a ghost writer has been used, this should be declared in the acknowledgements. I don’t think it’s appropriate that they should be listed as a main author if they have played no part in study design or conduct. However if it’s OK to acknowledge the statistician, why not do the same for the writer?
Why should professional expertise with the written word be an embarrassment?
About the author:
Les Rose is a freelance clinical scientist and medical writer, who most definitely was the author of this article.
For more information about his work and for queries please visit www.pharmavision-consulting.co.uk.
Is ghostwriting an acceptable practice for publishing clinical data?