Publication of clinical trial data: enough of the excuses!
Woodrow Medical Communications
I attended a MedComms Networking event last summer to listen to the thoughts of Ben Goldacre on the pharmaceutical industry. For those of you who don’t know Dr Goldacre, he is, according to his web page: “a best-selling author, broadcaster, medical doctor and academic who specialises in unpicking dodgy scientific claims from drug companies”. The briefing was fascinating. In particular, I was interested in Dr Goldacre’s arguments that the pharmaceutical industry are guilty of failing to publish all trial data (and in particular from negative trials), and that the industry fail to make enough data available for systematic reviewers and healthcare professionals to analyse. Dr Goldacre stated that this leads to publication bias and ultimately potential harm to patients (due to poor disclosure of adverse events) and unnecessary costs to us all (since the cheapest medications aren’t being used).
The audience, who were mostly made up of medical communications specialists involved in medical publications, were clearly somewhat shocked by Dr Goldacre’s talk and were ready to argue that the practices being discussed no longer occurred, that the industry now published all negative data, and that all trials are now adequately published (at minimum on clinical trial registries) from Phase II onwards due to legislation in the US and Europe.1,2 In reply, Dr Goldacre asked for evidence. However, aside from anecdotal comments, this did not seem at all forthcoming.
At the time, I agreed with the audience and felt that the bad practices had indeed generally ceased (i.e. that all negative data is now published). However, I also agreed with Dr Goldacre that the pharmaceutical industry should try to endeavour to publish ALL data from clinical trials, at minimum on clinical trial registries, and preferentially in peer reviewed journals. I wondered whether my colleagues working in medical publications were harmonised on these issues, and decided to set-up a survey to find out.
Working with Adam Jacobs, Jay Magrann and Lydia Dye-Stonebridge, Peter Llewellyn and Nigel Eastmond, we conducted a survey in August 2011 of members of the International Society for Medical Publication Professionals (ISMPP), the American Medical Writers Association (AMWA), the NetworkPharma community and other relevant groups on LinkedIn. Only professionals involved in developing, planning or publishing medical publications were eligible. Overall, 607 medical publication professionals from around the world responded and completed the survey. The majority of completers had more than 5 years of experience in medical publications, with over half involved in medical writing. We recently presented our findings at the ISMPP European meeting earlier in November 2011, and our poster won the “best original research” award. You can find a PDF here (we are also aiming to submit the data to a journal in the New Year).
“…a sizable proportion of respondents (32%) were aware of at least one piece of unpublished negative data from a clinical trial in the past 3 years.”
Surprisingly to me, a sizable proportion of respondents (32%) were aware of at least one piece of unpublished negative data from a clinical trial in the past 3 years (see Figure 1). Therefore, some negative data (whether that be from a subanalysis, a secondary trial endpoint, a Phase 1 study, or something else) are clearly not being published.
Figure 1. Awareness of negative data from ANY clinical trial not being published by a pharmaceutical company.
It is somewhat alarming to me that the pharmaceutical industry still do not publish all negative data. Some of the reasons given for non-publication are easier to understand than others, but none are acceptable in my view. The main reasons for non-publication over any time included:
• Compound discontinuation (40%)
• Journal rejection (36%)
• Poor trial design (31%)
• Lack of resources (27%)
• And damage to the product profile (27%).
Regarding ‘compound discontinuation’ and ‘poor trial design’, publication of negative data from these trials would help to ensure that other researchers do not waste time and resources making the same mistakes. ‘Journal rejection’ is no longer an argument with the huge breadth of journals now available, including some catering specifically for the publication of negative results. With respect to ‘lack of resources’, clearly if companies made publication of data more of a priority then this would no longer be an issue, as resources would be made available. Finally ‘damage to product profile’ is the most inexcusable of reasons and is clearly why Dr Goldacre and others are so quick to speak out against the industry.
“Anyone involved in medical publications should in my opinion be striving to publish all clinical data, even if this is just on a clinical trials website.”
We then went on in the survey to ask whether the pharmaceutical industry should be obligated to make trial data available from earlier time points (i.e. before Phase II). Overall, 36% of completers felt that the current time point of Phase II was acceptable. However, more (44%) felt that data should be available prior to Phase II, in line with my own opinion on this subject. When asked what media would suffice as making the data public, most deemed that www.clinicaltrials.gov, EudraCT, or a similar database would be adequate. Just over half indicated that the data should be published in a peer-reviewed journal. For what it is worth, Dr Goldacre’s opinion on this subject is that data should be published in entirety on a clinical trial registry, and also published in a journal.
Finally, we asked the participants what would be the main barriers to publishing all data from now onwards. The interesting opinions from the completers on this question are summarised in Figure 2. I wonder whether anyone of us could truly argue that these barriers are tolerable? In particular, would our friends or any member of our family for that matter accept ‘fear of data misinterpretation’ and ‘negative impact on products’ as being reasonable arguments?
Figure 2. What are the MAIN barriers to publishing all data from now onwards? (Top selections shown).
The pharmaceutical industry is far from an evil entity. Because of the pharmaceutical industry, our life expectancies are continuing to grow, many diseases are being slowly but surely eradicated, and many others are being effectively prevented through vaccination. In addition, medications provided by the industry improve the quality of life of patients and their families. As a worker of the industry, I am proud of these achievements. However, I am not proud of the bad decisions made by a small minority.
It is now time for us to stand up and be counted as individuals. Anyone involved in medical publications should in my opinion be striving to publish all clinical data, even if this is just on a clinical trials website. And none of us should find it acceptable to leave negative data unpublished… whatever the reason.
1. The FDA Amendments Act of 2007, Public Law 110-85.
2. EU pharmaceutical legislation (article 57 of the Regulation (EC) No 726/2004 and article 41 of the paediatric Regulation (EC) No 1901/2006).
Conflict of interest declaration: I am a medical writer, and I work on medical publications for pharmaceutical companies and other researchers.
About the author
Ryan Woodrow is a freelance medical communications consultant and medical writer with 15 years of experience in developing and implementing medical communication plans for the pharma industry. Ryan started his professional career within the medical information departments at Zeneca (now AstraZeneca) and Roche, where he was required to provide information to internal teams or to healthcare professionals on a number of drugs in written and oral format. Ten years ago, Ryan felt it was time for a new adventure and began looking for a career offering plenty of variety, but which gave him the opportunity to do more of the thing that he enjoyed doing most: medical writing, and he has not looked back. As well as writing and consulting, Ryan is actively engaged with assisting in the organisation of educational and networking events, and he is also a member of the ISMPP European Committee. Ryan can be contacted by email (firstname.lastname@example.org) or on Twitter (@woodrowmedical).
What can be done to aid the publication of all clinical data?