Reader Comments

Post a new comment on this article

The problem of selective publication will not be solved by refusing writing assistance

Posted by AdamJacobs on 06 Dec 2013 at 14:30 GMT

Many basic science researchers may not be familiar with the important role of medical writers in publishing clinical research studies. In her article on the role of the pharmaceutical industry in publications, Fugh-Berman [1] attempts to explain that role, but as professional medical writers, we are concerned that her explanation is biased and misleading.

Fugh-Berman’s article starts with a discussion of conflicts of interest, an appropriate place to begin our response. Readers will note that her own conflicts of interest include acting in court cases against the pharmaceutical industry, as well as directing PharmedOut, an organisation dedicated to campaigning against the pharmaceutical industry. To continue the legal metaphor, one might regard Fugh-Berman’s article as the case for the prosecution, rather than as an impartial academic discussion of the issues. We would like to present a more evidence-based point of view.

Fugh-Berman describes the problem of selective publication: not all research is published, and positive research is more likely to be published than negative research. This is indeed a problem, though the assertion that this is an “important tactic for industry” does not stand up to critical scrutiny. First, there is no evidence that the problem of selective publication is more widespread in industry-funded research than in academic research. Fugh-Berman cites a number of papers showing that many abstracts presented at conferences are never published in full, but none of those papers found that industry-funded research was less likely to be published than academic research. Second, her interpretation of the study by Bourgeois et al [2] is misleading. She describes results 2 years after publication, at which time industry funded studies were indeed less likely to have been published in that study, but Bourgeois et al’s longer term follow-up results showed that industry-funded studies were more likely than government funded research to be published eventually.

But in any case, it is unwise to rely on single studies: a systematic review should be the preferred form of evidence. Such a review by Schott et al in 2010 [3] found mixed evidence, but the majority of evidence showed that industry-funded studies were more likely to be published.

While incomplete publication is an important problem for clinical research, it is not limited to industry: it afflicts all clinical research.

Fugh-Berman also misunderstands the reasons why negative studies are less likely to be published, asserting that it is because of what authors submit, rather than what journals accept. In fact, those two things are interlinked: authors are more likely to submit papers if they believe there is a good chance of acceptance. So if authors submit fewer negative studies, that may be because journals accept fewer negative studies. Senn has shown that the available data is more consistent with selective publication being a function of journals’ reluctance to publish negative studies than a function of authors’ reluctance to submit them [4].

Thankfully, it probably is getting easier to get negative trials published now that journals are more aware of the problem of publication bias, but it would be a mistake to conclude that historical data on selective publication shows some kind of grand conspiracy: it is far more consistent with authors simply working within a flawed system not of their making.

Researchers both in academia and industry do need to make greater efforts to ensure that all their research is published. But how is this to be achieved?

Fugh-Berman warns against accepting statistical or writing assistance from research sponsors. She claims that it “provides an opportunity for results to be suppressed or spun”, but provides no evidence for that assertion. The prestigious JAMA recognises that changes have occurred in reporting clinical trials over the last decade and recently reversed its requirement for independent statistical analysis of industry-sponsored studies [5]. Moreover, we believe refusing assistance is entirely counterproductive. First, refusing statistical support is unrealistic. Many researchers are simply not qualified statisticians. But refusing writing assistance can also compromise the clarity, accuracy, and readability of the publication. Fugh-Berman mentions writing assistance only in the context of ghostwriting. A ghostwriter is someone who writes a paper, but whose name does not appear on the paper. We agree with Fugh-Berman that ghostwriting is unethical. She describes some instances where ghostwritten papers have been unethically written in a way that misleads. We join her in condemning such abuses in the strongest possible terms.

However, a casual reader of Fugh-Berman’s paper might be left with the impression that such abuses are common: an assertion for which there is no evidence. One might just as well argue that we must all avoid doctors when we are ill because Harold Shipman was a doctor and he murdered many of his patients [6].

In fact, there is a world of difference between the unethical ghostwriting that Fugh-Berman describes and ethical medical writing assistance. Professional medical writers who adhere to latest guidelines [7, 8] are emphatically not ghostwriters, as they declare their involvement in the paper and their funding source, and ensure that data are presented honestly. Research has shown that papers written with the assistance of professional medical writers are less likely to be fraudulent [9], more likely to comply with best practice reporting guidelines [10], and are accepted more quickly for publication [11].

The involvement of professional writers in preparing research for publication is common in clinical research, but perhaps less common in basic science. We would urge basic scientists not to be put off it by Fugh-Berman’s article. Writing assistance can free researchers from the task of writing up results, which many see as a chore. Provided researchers are careful to ensure that their writers follow best practice guidelines—one important aspect of which is that researchers must retain overall control of the content of the paper—there is no risk to the scientific quality of the publication.

As we have argued before, writing assistance may be the best practical means of tackling the problem of incomplete publication [12]. Refusing it will only make the problem worse.

Adam Jacobs, Cindy Hamilton, Art Gertel, and Karen Woolley, on behalf of the Global Alliance of Publication Professionals (http://www.gappteam.org)

References
1. Fugh-Berman A. How basic scientists help the pharmaceutical industry market drugs. PLoS Biology 2013;11:e1001716 (doi:10.1371/journal.pbio.1001716)
2. Bourgeois FT, Murthy S, Mandl KD. Outcome reporting among drug trials registered in ClinicalTrials.gov. Ann Intern Med 2010;153:158-66
3. Schott G, Pachl H, Limbach U, et al. The financing of drug trials by pharmaceutical companies and its consequences. Part 2: a qualitative, systematic review of the literature on possible influences on authorship, access to trial data, and trial registration and publication. Dtsch Arztebl Int 2010;107:295-301
4. Senn S. Misunderstanding publication bias: editors are not blameless after all F1000Research 2012, 1:59 (doi: 10.12688/f1000research.1-59.v1)
5. Bauchner H. Editorial Policies for Clinical Trials and the Continued Changes in Medical Journalism. JAMA 2013;310:149-150
6. Obituary: Harold Shipman. BMJ 2004;328:231
7. Jacobs A, Wager E. European Medical Writers Association (EMWA) guidelines on the role of medical writers in developing peer-reviewed publications. Curr Med Res Opin 2005;21:317-21
8. Graf C, Battisti WP, Bridges D, Bruce-Winkler V, Conaty JM, Ellison JM, et al. Good publication practice for communicating company sponsored medical research: the GPP2 guidelines. BMJ2009;339:b4330
9. Woolley KL, Lew RA, Stretton S, et al. Lack of involvement of medical writers and the pharmaceutical industry in publications retracted for misconduct: a systematic, controlled, retrospective study. Curr Med Res Opin 2011;27:1175-82.
10. Jacobs A. Adherence to the CONSORT guideline in papers written by professional medical writers. The Write Stuff 2010;19:196-200.
11. Bailey M. Science editing and its effect on manuscript acceptance time. AMWA Journal 2011;26:147-52.
12. Woolley KL, Gertel A, Hamilton C et al: Poor compliance with reporting research results – we know it’s a problem … how do we fix it? CMRO 2012:28:1857-1860

Competing interests declared: All GAPP members have or do hold leadership roles at associations representing
professional medical writers (eg, AMWA, EMWA, DIA, ISMPP, ARCS), but do not speak on
behalf of those organizations. GAPP members have or do provide professional medical writing
services to not-for-profit and for-profit clients

RE: The problem of selective publication will not be solved by refusing writing assistance

Fugh-Berman replied to AdamJacobs on 26 Dec 2013 at 20:29 GMT

Medical writers do serve an important role – but for industry, not research, as documented by medical writers themselves in peer-reviewed literature [1-4]. The American Medical Writers Association own survey found that every one of 61 respondents worked for industry, including pharmaceutical, medical device or biotech companies, medical education and communication companies, and contract research organizations (CROs) [5].

The medical writers belief that researchers should not refuse industry-sponsored writing and statistical assistance is understandable; as Upton Sinclair said “It is difficult to get a man to believe something when his salary depends on his not believing it.” But ethical researchers do their own analyses, rely on independent statisticians, and write their own papers.

Seventy percent of medical writers in AMWA’s survey said that academic “authors’” were not involved in industry-sponsored research articles until after data tables were generated. And despite the writers’ protests against ghostwriting, only half (49%) said their companies always acknowledge medical writers. [5]

I agree that incomplete publication affects clinical research in general, but the medical writers’ statement that “there is no evidence that the problem of selective publication is more widespread in industry-funded research than in academic research“ is wrong. A new study of 585 large randomized controlled trials registered in ClinicalTrials.gov found that 29% of 585 trials with almost 300,000 study participants remain unpublished and that trials that received industry funding were significantly less likely to be published (32%) than those without industry funding (18%) [6]. The Bourgeois analysis found similar results [7], as have systematic reviews [8,9].

Numerous studies show that industry funding affects not only publication, but the results of published studies, distorting evidence on benefits and risks. For a thoroughly referenced summary of these issues, see Kay Dickersin and Ian Chalmers’ "Recognising, investigating and dealing with incomplete and biased reporting of clinical research: from Francis Bacon to the World Health Organisation" [10].

In fact, the suppression of industry-funded trials launched an international movement, AllTrials, to demand the registration and publication of all clinical trials. I encourage PLOS Biology readers to join 61,969 colleagues in signing the petition at http://www.alltrials.net. Perhaps we need a similar petition calling for the publication of all basic sciences studies.

There are many errors and misleading statements in the medical writers’ letter. For example, they mistakenly assert that Bourgeois’ “longer term follow-up results showed that industry-funded studies were more likely than government funded research to be published eventually”; there was no significant difference in total publication rate between the two groups [7]. Schott’s review of how pharmaceutical companies influence drug trials, cited by the medical writers, does not support their assertions [11]. And they mischaracterize PharmedOut (http://pharmedout.org), which is a multidisciplinary research and education project of Georgetown University Medical Center that promotes evidence-based prescribing and educates healthcare professionals about pharmaceutical marketing practices.

I have long supported the hiring of medical writers by individuals, academic institutions, or government agencies to improve the readability of publications [12]. There is a huge difference, however, between correcting a researcher’s grammar and spinning a message for a corporation.


Adriane Fugh-Berman MD
Associate Professor
Department of Physiology and Pharmacology, and Department of Family Medicine
Georgetown University Medical Center

References

1. Fugh-Berman A, Dodgson S. The ethics of publication planning in the pharmaceutical industry. Open Medicine 2008;4(5): e40-43. http://www.openmedicine.c...

2. Braun SR. Promoting "low T": a medical writer's perspective. JAMA Intern Med 2013 Aug 12;173(15):1458-60. doi: 10.1001/jamainternmed.2013.6892

3. Logdberg L. Being the ghost in the machine: a medical ghostwriter's personal view. PLoS Med 2011 Aug;8(8):e1001071. doi: 10.1371/journal.pmed.1001071.
http://www.plosmedicine.o...

4. Matheson A. Corporate Science and the Husbandry of Scientific and Medical Knowledge by the Pharmaceutical Industry, Biosocieties 2008; 3 (4): 355-382.
http://www.law-lib.utoron...

5. Philips S. Authorship and Writing Practices in the Health Care Industry. AMWA Journal 2009; 24(1):4-8.

6. Jones CW, Handler L, Crowell KE, Keil LG, Weaver MA, Platts-Mills TF. Non-publication of large randomized clinical trials: cross sectional analysis. BMJ. 2013; 347:f6104. http://www.bmj.com/conten...

7. Bourgeois FT, Murthy S, Mandl KD. Outcome reporting among drug trials registered in ClinicalTrials.gov. Ann Intern Med 2010 Aug 3;153(3):158-66.

8. Lexchin J, Bero LA, Djulbegovic B, Clark O. Pharmaceutical industry sponsorship and research outcome and quality: systematic review. BMJ 2003;326:1-10. http://www.bmj.com/conten...

9. Sismondo S. Pharmaceutical company funding and its consequences: a qualitative systematic review. Contemp Clin Trials 2008 Mar;29(2):109-13.

10. Dickersin K, Chalmers I. Recognising, investigating and dealing with incomplete and biased reporting of clinical research: from Francis Bacon to the World Health Organisation. James Lind Library 2010: Available at: http://www.jameslindlibra....

11. Schott G, Pachl H, Limbach U, et al. The financing of drug trials by pharmaceutical companies and its consequences. Part 2: a qualitative, systematic review of the literature on possible influences on authorship, access to trial data, and trial registration and publication. Dtsch Arztebl Int 2010;107:295-301. http://www.akdae.de/Stell...

12.Fugh-Berman A. In reply to comments on The Corporate Coauthor. J Gen Int Med 2005 (October);20(10):972-973.

Competing interests declared: As reported in the original article, I direct PharmedOut, a Georgetown University Medical Center research and education project that advances evidence-based prescribing and educates healthcare professionals about pharmaceutical marketing practices. I am a paid expert witness on behalf of plaintiffs in litigation regarding pharmaceutical marketing practices, including litigation against Pfizer (which owns Wyeth), GlaxoSmithKline, and Takeda Pharmaceuticals.

RE: RE: The problem of selective publication will not be solved by refusing writing assistance

AdamJacobs replied to Fugh-Berman on 08 Jan 2014 at 16:27 GMT

We invite Fugh-Berman and others who are unclear about the role of medical writers to carefully examine the evidence and to choose large, recent studies over older publications and anecdotal experience. Taking the time to read beyond the abstract and delve into details within the body of publications reveals a number of misinterpretations by Fugh-Berman that merit clarification. We also offer practical tips for those who are concerned about how to work ethically with a medical writer.

Fugh-Berman starts by describing a small survey by Philips as “the American Medical Writers Association own survey”. For a start, this was not AMWA’s own survey, it was simply a survey that happened to be published in AMWA’s journal. But more importantly, the fact that all respondents worked for industry is hardly surprising, given that the survey was specifically of medical writers working for industry. That finding was purely a result of the design of the survey, and tells us nothing about the proportion of medical writers who work for industry.

We are puzzled as to why Fugh-Berman sees an apparent problem with the lack of involvement of authors until after data tables were generated. Generating data tables is the job of statisticians, not of clinicians. It is hardly meaningful for clinicians to start considering how to publish results until the results are available, and in practice that means until after data tables have been created. What is unethical is if authors are not involved until after the first draft had been written, which was reported by only 5% of respondents. Should it have been 0%? Of course, but 5% of respondents reporting too late an involvement of authors is very different to 70%.

The statement that “only half said their companies always acknowledge medical writers” is misleading, as that’s not what the survey actually asked. It asked whether companies had policies to acknowledge medical writers. Many companies simply had no policy at the time of the survey in 2007, but it does not therefore follow that they did not acknowledge medical writers in practice.

Perhaps it would be worth looking at what the real AMWA survey (conducted jointly with the European Medical Writers Association) found [1]. This survey, which was based on a much larger sample than the one by Philips, was repeated in 2005, 2008, and 2011, and found that the prevalence of unacknowledged contributions decreased substantially during that time. The most recent figure, from 2011, showed that 33% of medical writer’s contributions were unacknowledged. Of course, 33% is too high: it should be zero. This, however, is not a good reason to refuse medical writing assistance. We would argue, and I hope Fugh-Berman would agree with us, that the important thing is that those authors who work with professional medical writers ensure that their writers are acknowledged, so that we can continue to fight against ghostwriting.

Fugh-Berman disputes our statement that “there is no evidence that the problem of selctive publication is more widespread in industry-funded research than in academic research”. In support of this, she cites a recent study by Jones et al (2013), which she claims shows industry trials were significantly less likely to be published. That is true, but it’s not the full story. Some of the unpublished trials had results posted on clinicaltrials.gov, so were still disclosed in the public domain. When those are added to the number of publications, the difference between industry and non-industry studies is no longer statistically significant (P = 0.16, Fisher’s exact test).

She also claims that the Bourgeois study found similar results. As we explained in our previous response, that is simply not true. Although Bourgeois et al did find that industry studies were less likely to be published at 24 months after trial completion, they were more likely to be published eventually. We note that the abstract of that paper only reports the results at 24 months, so perhaps Fugh-Berman did not read the full paper? Table 3 of Bourgeois et al’s paper shows that 230/346 (66%) industry-funded studies were published, and 132/200 (66%) of non-industry studies were published (combining government-funded with nonprofit or nonfederal funded into a single category). If you add trials with results disclosed on clinicaltrials.gov, the disclosure rate increases to 305/346 (88%) for industry studies, but remains at 132/200 66% for non-industry studies. That difference is highly statistically significant (P = 0.000000001, Fisher’s exact test).

Fugh-Berman also cites 2 systematic reviews finding that industry-funded studies were less likely to be published. One of those (Lexchin et al) is from 2003, so is less relevant than the more up-to-date review by Schott et al that we cited. The other (Sismondo 2008) is, despite the title, not actually a systematic review, as it did not include systematic searches of literature databases.

We are puzzled as to why Fugh-Berman thinks the more up-to-date review by Schott et al does not support our assertions. We stated that it “found mixed evidence, but the majority of evidence showed that industry-funded studies were more likely to be published”. This information wasn’t presented in the abstract: is this perhaps another case of Fugh-Berman failing to read the whole paper? Here is what Schott et al said:

“There are varying data on whether the type of funding of a trial influences the likelihood of publication. While an investigation of protocols provided to an ethics committee showed that funding independent of pharmaceutical companies is associated with a greater likelihood of publication (e20), two studies in oncology (e21, e22) and an investigation of the protocols supplied to a different ethics committee (e23) revealed that trials funded by pharmaceutical companies were published more often (e22, e23) or more quickly (e21). A further investigation of publications of registered trials showed that low publication rates are not limited to particular sponsors of a trial; rather, they are a problem with both industry-funded and state-financed trials (e24).”

In other words they found 3 studies showing higher rates of publication from industry, 1 showing higher rates from academia, and 1 showing no difference. We stand by our assertion that this shows mixed evidence, but with the majority of evidence showing higher rates of publication from industry.

Fugh-Berman states that there is a huge difference between correcting a researcher’s grammar and spinning a message for a corporation. We agree. There is, however, no evidence that medical writers contribute to data being spun. Studies that have attempted to quantify spin have failed to find a difference between industry and non-industry studies [2, 3]. We are not aware of any studies looking specifically at the association between medical writing assistance and spin, but we reiterate our earlier remarks that studies have shown use of professional medical writers is associated with a decreased risk of fraud, which might perhaps be seen as an extreme example of spin. However, for any authors who remain worried that using a professional medical writer will introduce spin, the solution is simple: ensure that your agreement with the writer makes clear that you control the content of the paper, which any professional medical writer following published reporting standards would expect anyway. If you believe anything in the paper represents spin, simply ask the medical writer to change it.

It is also worth pointing out that a good medical writer does much more than simply correcting a researcher’s grammar, such as educating researchers about best-practice reporting guidelines and ensuring that data are presented clearly, accurately, and with fair balance.

We are all agreed, I believe, that incomplete publications are a problem. We agree with Fugh-Berman that the All Trials initiative has laudable aims, and that an equivalent initiative for basic science would be welcome. However, admonishing people to publish is unlikely to achieve anything by itself: we also need to give researchers practical help in ensuring that research gets published, and that the resulting publications meet the highest standards of scientific integrity

Professional medical writers provide that help. We hope that basic science researchers will realise that, in the fight against hidden data, professional medical writers are part of the solution, not part of the problem.

Adam Jacobs, Cindy Hamilton, Art Gertel, and Karen Woolley, on behalf of the Global Alliance of Publication Professionals (http://www.gappteam.org)

References

1. Hamilton CW, Jacobs A. Evidence of decreased ghostwriting in a series of three surveys of medical communicators conducted between 2005 and 2011 (abstract). AMWA J. 2012;27(3):115

2. Boutron I, Dutton S, Ravaud P, Altman DG. Reporting and Interpretation of Randomized Controlled Trials With Statistically Nonsignificant Results for Primary Outcomes. JAMA. 2010;303(20):2058–2064. doi:10.1001/jama.2010.651

3. Vera-Badillo FE, Shapiro R, Ocana A, Amir I, Tannock IF. Bias in reporting of end points of efficacy and toxicity in randomized, clinical trials for women with breast cancer. Annals of Oncology 24: 1238–1244, 2013. doi:10.1093/annonc/mds636

Competing interests declared: As stated previously

RE: RE: RE: The problem of selective publication will not be solved by refusing writing assistance

Fugh-Berman replied to AdamJacobs on 18 Jan 2014 at 02:10 GMT

The four authors comprise 80% of the grandly named Global Alliance of Publication Professionals http://www.gappteam.org/i... a five member rapid-response public relations team apparently set up to attack anyone critical of pharma-funded writers. In their own words, GAPP’s first aim is to, “Provide a timely and credible response to influential stories about medical publication professionals (eg, professional medical writers, publication planners).”

GAPP’s "response criteria" states that it will respond to stories that “appear in an influential information source (influential in terms of prestige, geographic reach, likely readership; could be a newspaper, journal, social media platform)” that “lack information or include misinformation about publication professionals."

As stated on its website, "GAPP members will aim to respond quickly (within 24-48 hours),” so set your watches. I have requested that PLOS deny this front group unfettered access to this forum; in any case, I do not intend to carry on further correspondence with industry-paid professionals.

Adriane Fugh-Berman MD
Department of Pharmacology and Physiology and
Department of Family Medicine
Georgetown University Medical Center


Competing interests declared: As reported in the original article, I direct PharmedOut, a Georgetown University Medical Center research and education project that advances evidence-based prescribing and educates healthcare professionals about pharmaceutical marketing practices. I am a paid expert witness on behalf of plaintiffs in litigation regarding pharmaceutical marketing practices, including litigation against Pfizer (which owns Wyeth), GlaxoSmithKline, and Takeda Pharmaceuticals.

Updated Statement of Competing Interests for Adam Jacobs and colleagues

pbio replied to Fugh-Berman on 25 Feb 2014 at 11:36 GMT

Regarding the initial and subsequent comments posted by Adam Jacobs and colleagues, beginning 6th December 2013, comprehensive competing interests statements for all members of the group can be found below:


Adam Jacobs PhD MSc FICR CSci Cert IoD runs " a growing biometrics Contract Research Organisation providing expert medical writing, statistical consultancy, and data management services." (http://uk.linkedin.com/in....) He owns Dianthus Medical Limited and has been Director since 1999. Dianthus Medical Limited provides "medical writing, data management, and statistical services, mostly to pharmaceutical companies." Previously was a medical writer for Medical Action Communications and Hammersmith Medicines Research. Current press officer and former exec committee member of European Medical Writers Association (EMWA), where he "formed and led EMWA’s ghostwriting task force, responsible for development and publication of guidelines on the role of medical writers in peer-reviewed publications." CV here http://www.esforum.eu/201...

Cindy Hamilton PharmD ELS was president of American Medical Writers Association (AMWA) and current principal of Hamilton House in Virginia Beach, "established in 1990 to deliver medical and scientific communications to the pharmaceutical industry, medical profession, and other scientific communities." CV here http://www.hamiltonhousev...


Art Gertel BS MS was also a president of American Medical Writers Association (AMWA), and is a Fellow of AMWA and EMWA and on The International Publication Planning Association (TIPPA) Advisory Board and the Advisory Board for Hummingbird IRB. He is Vice President, Regulatory and Medical Affairs for TFS, Inc. in NJ. He is also a Senior Research Fellow with the Centre for Innovation in Regulatory Science (CIRS), a London-based “Think Tank” dedicated to improving Benefit-Risk evaluation and decision-making in the development and approval of new medicines. He serves as Co-Chair of the Global Ethical & Regulatory Innovation (GERI) task force, under the auspices of the Alliance for Clinical Research Excellence and Safety (ACRES) "With more than 30 years of experience in the pharmaceutical industry, Art has had a broad range of responsibility in virtually all phases of pharmaceutical drug research and development....Art’s prior experience includes heading the Medical Communications Department at Schering-Plough, with responsibility for preparing all new drug applications, worldwide; directing the Medical and Scientific Communications Department at Hoffmann-LaRoche" Bio here http://www.beardsworth.co...


Karen Woolley BHMS Ed Hons PhD CMPP was president of ARCS (Association of Regulatory and Clinical Scientists Australia Ltd) Australia, is an Honorary Fellow of AMWA, and Director of The International Society for Medical Publication Professionals (ISMPP). She is CEO of ProScribe Medical Communications which "prepares high quality, scientifically robust communications that ethically inform our target audience about the safety and efficacy of therapeutics." As of January 2014, ProScribe Medical Communications has become part of the Envision Pharma Group (www.envisionpharmagroup.c...). "Karen has spent the last 25 years immersed in medical writing and clinical research within academia and the pharmaceutical industry." Bio here http://www.proscribe.com.....

No competing interests declared.