Industry influence on the published record
A sponsor's reach into a study and its reporting is often larger than the published paper shows. Critics, writers' associations, editors and medical writers have disputed how far that reach shapes the published record, what a company does with results that cut against its product, whether journals should publish industry-funded research at all, and how far promotion has spread into medical communication.
The sponsor's part in research and its reporting
Industry sponsors the vast majority of clinical research. A thematic study found bias in every one of 12 themes, among them the research question, the choice of dose or comparator, trial design and the reporting of results, and concluded that this may undermine confidence in such research. The study did not systematically compare industry-supported with other studies, nor say which biases were commonest. Industry's investment in research, and its documentation by medical writers, helps deliver life-saving medicines but can also distort evidence to favour and protect the investment.
A number of articles have suggested that industry-sponsored trials usually favour the company's product, and this has raised concern about bias and influence in industry-sponsored studies. One study examined sponsor involvement in the conduct and reporting of randomised clinical trials published in The Lancet in 2008 and 2009 by comparing each publication with its protocol, which the journal had requested with manuscripts since 2002. Sixty-nine of the trials were industry-sponsored and twelve were industry-funded but independently conducted.
In the majority of cases the sponsor or a contract research organisation was involved in reviewing and verifying the information in case report forms, in data entry, data storage, data analysis and publication of the results, rather than these tasks being done independently by academic authors. Only two trials had a completely independent analysis. Even among the twelve independently conducted trials, the sponsor seemed to have some influence over the conduct of the trial or the reporting of the results. Medical writing assistance from the sponsor or from someone the sponsor hired was described in 37 of the studies (54%). The study's authors suggested that it may be the responsibility of journals to insist on more transparent reporting of the sponsor's role in processes such as data processing, statistical analysis and report writing, and that all journals should consider asking for study protocols and raw data to be submitted and for independent data analysis.
Authors should never allow a pharmaceutical company or other research sponsor to do the statistical analysis of their data.
It is well recognised that financial conflicts of interest can affect how study results and conclusions are presented. Pressure to publish and commercial pull can push academics to lend their names to papers they did not contribute to. Industry-sponsored articles make up between 67% and 75% of those published in The Lancet, The New England Journal of Medicine, JAMA and Annals of Internal Medicine, and such sponsorship can influence, directly or indirectly, the people involved in writing, editing or publishing scientific information. A study of health researchers' attitudes and experiences found guest authorship widespread, at 65% (101 of 155) in India, along with a lack of awareness of conflicts of interest and how they influence research and conflicting views on whether commercial links to companies would influence researchers.
Not reporting negative findings wastes time and resources, biases meta-analyses and produces false conclusions of benefit, which bring science into disrepute and may harm patients. John Ioannidis showed in 2008 how an illusion of antidepressant effectiveness was created by many small randomised trials with clinically non-relevant outcomes, short follow-up, biased study populations and incorrect interpretation of statistical significance. The conclusion has been intensely contested, but a later analysis of all published meta-analyses, including one of antidepressant data submitted to the Food and Drug Administration for approval, supported it, concluding that most of the benefit is due to placebo response, with no clinically meaningful difference between drug and placebo. Marcia Angell, a former editor of The New England Journal of Medicine, holds that much published clinical research cannot be trusted and that even trusted physicians and guideline recommendations cannot be relied on.
Pressure to favour an industry has reached even the abstracting of the literature: one writer who abstracted and indexed articles for the Council for Tobacco Research in New York before 1974 recalls being expected not to abstract whole articles but to extract the comments favourable to tobacco.
Ghost management
There are many reports of medical journal articles being researched and written by or on behalf of pharmaceutical companies and then published under the names of academics who had played little role earlier in the research and writing. In extreme cases, a company pays contract research organisations to run trials, analyses the data in-house, has professionals write the manuscripts, asks academics to serve as their authors, and pays communication companies to shepherd them through publication in the best journals. The resulting articles affect the conclusions found in the medical literature and are used in promoting drugs to doctors. As The New York Times reported, an Annals of Internal Medicine article on Merck's ADVANTAGE trial of Vioxx omitted some trial participants' deaths, and its first author, Jeffrey Lisse, said that Merck had designed, paid for and run the trial, had come to him after the study was completed, and had written the initial paper before sending it to him for editing.
Sergio Sismondo's September 2007 article in PLoS Medicine enlarged the focus from ghost writing to ghost management: pharmaceutical companies and their agents controlling or shaping multiple steps in the research, analysis, writing and publication of articles. Such articles are ghostly because the signs of their actual production are largely invisible, the academic authors named at their head giving corporate research a veneer of independence and credibility; and they are managed because the companies shape the message conveyed by an article or by a suite of articles. On Sismondo's account a substantial percentage of medical journal articles are ghost managed, which allows the pharmaceutical industry considerable influence on medical research and makes that research a vehicle for marketing. One industry statement gives the purpose of data as supporting, directly or indirectly, the marketing of the product.
Ghost writing and honorary authorship are not in themselves scientific problems, Sismondo argues, though they become so when they shape science to meet particular interests. Some honorary authors are senior professors and department chairs added because of local academic politics rather than at a company's request. Some busy independent research units hire writers to improve manuscripts, and one commentator has argued that professional medical writers can benefit the scientific community when used in a responsible manner. The writing of a manuscript may not be the key point at which behind-the-scenes influence is exerted: study design, statistical analysis or the choice of where to place a manuscript may be equally important. It has been repeatedly and firmly established that pharmaceutical company funding strongly biases published results in favour of the company's products. Ghost management, Sismondo argues, amplifies that bias, because when one set of commercial interests exerts influence at several stages of research, writing and publication, it shapes the resulting article; the bias in turn affects medical opinion and practice, and ultimately patients.
Because ghost management is hidden, how common it is cannot be told from published exposés. Current practice legitimately allows people to be authors on the basis of narrow contributions, so many near-honorary authors find little reason to feel uncomfortable with their roles, and fully honorary authors may not see enough of how their articles were produced to know they were ghost managed. It is in the interest of none of the writers, authors, sponsors or their agents to reveal the process, so a number of the published accounts have stemmed from legal proceedings and investigative journalism. Surveys quantifying ghost writing do not measure ghost management, because management may not involve writing and writing may not be managed. Information about ghost authors, people who should receive author credit, nevertheless strongly suggests that ghost management is common. A study comparing protocols with publications for industry-initiated trials approved by the Scientific-Ethical Committees for Copenhagen and Frederiksberg in 1994–1995 found evidence of ghost authorship in 75% of the publications (95% confidence interval 60% to 87%); company statisticians were common unacknowledged contributors, as were the creators of trial designs and protocols and the writers of manuscripts. The same study found that most of the trials whose protocols had been submitted, 172 of 274, were never begun, completed or published.
On Sismondo's account, ghost management of medical journal publications is a substantial business employing thousands of marketers, writers and managers. Besides the publication planners, a much larger number of medical writing companies and individual writers create articles and presentations without broader publication planning, and these may be adjuncts to the planners. Several publication planning firms are owned by major publishing houses: Excerpta Medica describes itself as an Elsevier business whose relationship with Elsevier gives access to editors, editorial boards and opinion leader networks, and Wolters Kluwer Health points to its publisher Lippincott Williams & Wilkins and to Ovid and its other medical information providers. At least three of the world's largest advertising agencies own both medical education and communication companies and contract research organisations. The business is large enough that the industry has established the International Publication Planning Association, which appears to be dominated by pharmaceutical companies and organises meetings, keeps a directory of experts and gives awards to planners. The International Society for Medical Publication Professionals also organises meetings, has policy committees and posts job advertisements. Both associations compete with for-profit companies offering similar services, such as the Center for Business Intelligence, which held forums on strategic publication planning in 2005 and 2006. Sismondo sees no straightforward solutions short of large changes to medical publishing or research that would sequester pharmaceutical company funding from research and publishing or from marketing, and until such changes come about hopes at least for more awareness of and responsiveness to the issue.
AMWA replied to the article on 3 October 2007 in a response posted online at PLoS Medicine. The reply contested the statement that the association's membership comprises mostly employees of medical education and communication companies, which it said extrapolated from the affiliations of two of five volunteer officers in a 5,500-member organisation whose members work in pharmaceutical and device companies, universities and medical schools, hospitals, non-profits, government agencies and journals, with fully a third self-employed. It also contested the statement that the association's conference is dominated by publication planning, when none of the 250 educational lectures, workshops and roundtables at its forthcoming national conference was on that subject and very few included topics even related to such companies. It noted that the article cited neither the association's Code of Ethics nor its position statement on the contributions of medical writers, which call for routinely acknowledging the pertinent professional and financial relationships between all authors and sponsors and the substantial contributions of medical writers to manuscripts, and that it cited neither Good Publication Practice nor the ICMJE guidelines, both of which the association endorses and integrates into its educational programme. The association's position statement holds that the involvement and proper acknowledgment of medical writers is legitimate and ethical. The reply hoped that the mischaracterisation reflected only inadequate research and not outright bias, and agreed wholeheartedly with the article's contention that scientific publications must be kept as free from bias as possible. It carried a competing-interests statement of its own, its lead author being employed by one of the companies the article had cited as a medical education and communication company.
A company's control over unfavourable results
In one writer's account, the road to publication of one study revealed that companies may be more concerned with protecting profits than with publishing facts. In 2014 The New York Times reported on court documents released in several lawsuits in which families of deceased patients claimed that Boehringer Ingelheim, the maker of the anticoagulant dabigatran, marketed as Pradaxa, had failed to inform them properly of the risk of taking it. Approved in 2010, the drug had earned more than $2 billion in US sales and been prescribed to more than 850,000 patients there, gaining substantial ground on warfarin. It had also been linked with multiple cases of fatal bleeding, with more than 1,000 deaths reported, and with gastrointestinal problems and an increased risk of heart attack and of haemorrhage; the European Medicines Agency identified bleeding as the most serious side effect, occurring in 1 of 10 patients. The company stood by the drug, pointing to its backing by the Food and Drug Administration and by multiple clinical trials.
A report led by the company's own clinical programme director found that not all patients, and in particular not all older patients, metabolise the drug in the same way, and that a small population would benefit from monitoring their blood: a small number of patients did not absorb the drug efficiently, while others absorbed it a little too well, raising their risk of bleeding. The original report went on to describe the ideal blood level of the drug, saying that keeping patients within that range would be optimal for preventing both stroke and bleeding. The finding weakened one of the drug's biggest selling points against warfarin, that blood tests are not a prerequisite for using it.
The documents, which included emails, memos and internal presentations, showed employees at different levels questioning the plans to publish the report. Their main concern was how the results might negate a decade's worth of work proving that patients taking the drug would not need regular tests, and concern was also raised that the results would not help in the company's race against other new anticoagulants such as Xarelto and Eliquis. The paper was published in the end, but not with all of its original results: according to a company representative, the scientist decided not to report the suggested optimal dosage. The company said the documents represented small fragments of the robust discussion and debate that is a vital component of all scientific inquiry and of the research and development of any important medication.
In the same writer's view, the released evidence suggests several questionable internal practices from both legal and ethical standpoints, and the controversy focuses on whether pharmaceutical companies can be trusted with handling their own results. The outcome was positive in that the results were seen outside the company, and the company would have to rethink a strategy that sold the drug as a one-size-fits-all product requiring no testing; the difficult path to publication nevertheless suggests a worrying underlying problem. On this account it falls to regulators to question the legal and ethical integrity of large pharmaceutical companies and to decide whether they need a hand in weighing profits against safety, even where the safety issues involve only a small number of patients.
Whether journals should publish industry-funded research
The BMJ and its associated journals stopped publishing research funded by the tobacco industry, on the grounds that the research is corrupted and the companies publish it to advance their commercial aims, oblivious of the harm they do. A 2014 head-to-head debate in the BMJ asked whether the same arguments apply to research funded by the drug industry, so that journals should also stop publishing drug company-funded trials.
The affirmative case claims that drug company-funded research is flawed and published to encourage sales, and proposes a new model: trial planning should begin with a systematic review of previous work to determine whether a new trial is necessary; if it is, the systematic review and the new trial protocol should be posted publicly on the internet for review and comment; the statistical analysis plan should be written before any data are available for analysis and posted with the protocol; on completion, the entire anonymised dataset should be made available for anyone to analyse; and journals should then publish the results of the systematic review and of all independent analyses of the trial data.
The negative case holds that the tobacco and drug industries are fundamentally different, since tobacco products harm health while pharmaceutical products aim to improve it, and that there are plans to increase integrity in the publication of drug company-funded research. It points to the many steps being taken to improve transparency in the evidence base for new drugs, such as mandatory prospective trial registration, reporting of all results and access to patient-level data on the benefits and harms of interventions, and holds that those rules should also be applied retrospectively to previously unreported trials. The BMJ is keen to publish papers from the Restoring Invisible and Abandoned Trials initiative, in which academics find and publish previously unreported trials whose original investigators declined to publish, and also trials that showed no evidence of benefit, provided the research questions are important and the methods robust.
The debate closes by asking whether editors would be afraid or unable to ban drug company-funded research, given the income journals receive from the pharmaceutical industry through advertising, reprints and sponsorship. The BMJ's editor-in-chief had stated that if these efforts did not soon bring about a necessary sea change in the way industry-funded trials are performed, the journal might well decide to stop publishing them, and that whether an editor would survive such a decision was a question she might have to test.
Journals have a bias against medical writers who touch anything to do with industry, William L. Lanier, editor-in-chief of Mayo Clinic Proceedings and professor of anaesthesiology at the Mayo Clinic, told medical writers in 2010. Mayo Clinic Proceedings took a different tack, set out in an editor's note published in 2006 and still in force in 2010: it could not adhere to its mission without publishing industry-funded research.
In that note Lanier wrote that, while there had occasionally been problems with drug industry-sponsored manuscripts, the overwhelming majority of problem authors and manuscripts in his recollection of more than seven years had not involved an identifiable funding source, drug companies included, as the cause of the aberrant behaviour; the moral lapses most commonly appeared to be related to ego and greed and to authors' desire to acquire or protect academic or clinical turf.
Lanier credits industry support of research with bringing capital, systems and facilities to conduct influential high-quality research, and holds that many of the brightest minds in biomedical research are employed by or affiliated with industry. Competition for market share and to get products into the marketplace, he adds, can also lead to shortcuts, biased research designs, intentional deception and less-than-ideal quality control, all of which can contaminate the medical literature and affect its overall quality. In his view the better elements of industry do not need reforming and are doing a great job, but the worst elements show little interest in reforming, so external forces must be brought to bear.
On Lanier's account, investigators and authors of industry-sponsored research bring considerable experience, access to critical patient populations and teams of collaborators, and a near-comprehensive understanding of the important arguments of the day. Incentives can cloud their judgement and hinder their ability to critique a proposal, and the incentive is usually not money but having their name associated with something important that may change health care; the threat of being "outed" by industry may drive them toward somewhat questionable practices they would not otherwise consider. Payment from industry does not make the resulting work illegitimate: Lanier did drug company studies because the companies had the cerebroprotective drugs, and not a single grant ever paid what the research cost.
Some editors, Lanier held, can sometimes confuse the roles of the journal with their own personal agendas. He pointed to two editors at two prominent medical journals who between them had more than 225 publications, nearly all editorials, commentaries or letters, much of it about industry-related conflicts, with no original scientific publications of their own, and held that journals, like corporations, need oversight panels. When the system fails, on his account, the cost is fewer critical high-quality articles in print, fewer new drugs and devices approved, fewer treatment options, less competition and innovation, fewer journal reviewers, less advertising, less need for medical writers and authors, and a lack of trust by those served. What is needed, he said, is détente: guidelines and regulations on publishing and clinical research need to be simplified with more common sense, or the unintended consequence of an unbridled increase in regulation may be the suffocation of clinical research and discovery.
Promotion in medical communication
Medical writing communicates research findings to the medical community and the public, and that relationship is strengthened when results are communicated in a transparent, timely and unbiased manner. A pertinent question is whether promotion is stealthily and steadily infiltrating every element of medical communication, tilting the balance toward an unhealthy reporting bias. Promotion is imperative for the pharmaceutical industry to convey important product information and encourage judicious use, but an unhealthy tilt in the balance between promotion and science can threaten patient safety. The balance can be ensured by following best practices in reporting research findings and adhering to regulatory guidance, and medical writers, as a critical touchpoint between medical practitioners and industry, can play an important role in keeping it.
The World Health Organization defines drug promotion as all informational and persuasive activities by manufacturers and distributors to influence the prescription, supply, purchase or use of medicinal drugs. Promotion can dominate marketing literature aimed at physicians, medical journalism, slide kits, white papers, industry-sponsored meetings for key opinion leaders and sponsored websites, and it can also creep into regulated documents and industry-funded journal articles. Continuing medical education may be commercially funded, but providers accredited by the Accreditation Council for Continuing Medical Education ensure that its content is fair, balanced, practice-based and independent of commercial influence, with regular audits for compliance. Medical writing can tilt toward promotion when messages highlight a product's benefits and underplay or even hide its risks. Manipulation of research findings through trial design, dose selection, choice of comparator and biased reporting, publishing positive results and suppressing negative ones, has been documented. Unregulated promotion on social media poses a further challenge, and promotional literature has been reported to influence physicians' prescribing significantly. Medical writers shoulder the responsibility for the right balance of science and promotion as more and more of the writing is outsourced by sponsors to medical writing service providers.
Promotion can also reach a journal as an advertorial, a piece structured as an original article whose content is overtly promotional. Four writers who objected to one such piece in the AMWA Journal in 2007 held that running what is essentially a free advertisement in a publication that accepts paid advertising violates good journalistic standards and risks alienating those who pay for their advertisements to appear, and that an original article on the same topic would have described similar companies and resources and what such companies do in general. Noting that most medical journals do not accept advertising disguised as education, the writers asked the journal to consider developing a policy on publishing advertorials.
Information given to the medical community and patients needs to be grounded in reality, without false claims. Public harm has followed when important risks associated with a drug did not reach regulators and the medicine stayed on the market, and promotional pressures are reported to cause even large and ethically inclined companies to brush aside safety risks. One commentator holds that a veil of independent research spending in science biases studies and their results toward their funders, citing journals that claim to perform peer review they do not perform and a National Institutes of Health study, funded by an alcohol lobby, that claimed cardiovascular benefits for drinking. Commercial interactions between the pharmaceutical industry and the medical profession can dilute the responsible use of medicines, and the resulting conflict of interest can undermine patients' trust.
Keeping science ahead of promotion
Medical writing conveys information that can change physicians' prescribing and patients' lives. Good papers can persuade and inspire for the better, but persuasion without sound scientific rationale can skew readers' understanding, so medical writers walk a tightrope between academia and commerce. To prevent bias, the writer should start with an open mind and an awareness of the ethical principles and reporting guidelines. A comprehensive literature search for all relevant evidence, positive or negative, lays the foundation for an unbiased perspective. The writer then sifts the data and synthesises it clearly into an accurate picture, presenting technical, difficult data in very simple language so that the reader can interpret it accurately, because obscure or ambiguous writing can leave readers confused and unable to make adequately informed decisions. Working closely with academic investigators, sponsors, journal editors and reviewers can put the writer in a unique position to practise total transparency in reporting evidence without partisanship. Academic authors would do well to follow the Chinese scholar Lu Ji, who wrote that writing is a struggle between presence and absence and urged writers to weigh each word on a scale and use a measuring cord to make their cuts.
Registering and publishing all clinical trial results, postmarketing safety and effectiveness information included, increases transparency and avoids the publication bias that the quiet burial of negative results often imposes. Negative results should be considered as important to publish as positive ones, so that drugs stay in use only if they have the right benefit–risk balance, without magnified effectiveness or underestimated harm. A virtual firewall between the people doing the research or analysis and the sponsors, or a funding source independent of the pharmaceutical sector, can counter commercial bias, and disclosure of all funding of health professionals and patient groups helps make those interactions transparent. A writer may not be able to create a firewall or enforce disclosure, but awareness of such measures helps the writer take a more balanced approach.
Ethical drug promotion contributes to the quality of health care through the rational use of medicines. It allows an unbiased flow of accurate, evidence-based information among pharmaceutical companies, health professionals, researchers and regulators; transparent reporting of clinical trial data enables life-saving drugs and devices to reach the market quickly; balanced communication of a product's risks and benefits allows unsafe drugs to be withdrawn in time; and credible information about treatment options and scientific advances lets health professionals and patients make informed, judicious decisions. The medical writing community has an obligation to shape scientific understanding ethically, and clarity, objective interpretation of clinical results, adherence to guidelines and ethics, prompt disclosure of conflicts of interest, and not being swayed by pressure from collaborators are key levers for minimising promotion and maximising science. Balance between science and promotion is achieved when all key stakeholders accept that transmitting objective and unbiased science is the only goal of medical writing.