Jane Feinmann: Is the current system of publishing clinical trials fit for purpose?

jane_feinmannThis question was the title of a meeting of the Medical Journalists’ Association last week, and, perhaps surprisingly for an audience made up almost exclusively of medical journalists, the response was a resounding no.

So what happened? Medical journals, the main vehicles for publishing clinical trials today, are after all the “gatekeepers of medical evidence”—as they are described in Bad Pharma, Ben Goldacre’s 2012 bestseller. They have a robust, evidence based reputation for usefulness: publishing research that translates scientific discoveries into practical applications—and that includes negative results. “In recent times, one in three trials published in The BMJ have been of a negative result,” Dr Trish Groves, deputy editor of The BMJ and editor in chief of BMJ Open, told the meeting. “If there is bias, it’s more likely to stem from the academics themselves who don’t submit negative research because they see it as dull.”

Thus Bad Pharma, with its uncompromising subtitle, How Drug Companies Mislead Doctors and Harm Patients, identifies the pharmaceutical industry as almost solely responsible for the dangerous and longstanding practice of cherry picking research for publication—with one in two trials going unpublished, and positive studies twice as likely to be published as those with negative results. And the great and the good in medicine, it seems, feel the same way. The Alltrials campaign, launched two years ago on the back of Goldacre’s book, has attracted an extraordinary level of support. GlaxoSmithKline was an early supporter, alongside 250 other medical organisations including the National Institute for Clinical and Health Excellence, the Medical Research Council, the Wellcome Trust, all the Royal Colleges, and The BMJ.

But this apparent consensus was challenged at the MJA meeting by, among others, Professor Stephen Senn, a former pharmaceutical statistician at University College, London and Glasgow University, and now a consultant to the pharmaceutical industry and head of the Competences Center for Methodology and Statistics at CRP-Sante (a publicly funded healthcare research institute) in Luxembourg. Professor Senn has long argued the AllTrials case, he insisted. “There’s no doubt that obtaining a license to market a drug should involve an obligation to share the results with interested parties,” he said.

His point, however, was that this sharing should not involve medical journals. “The medical press must become irrelevant to publication of clinical trials.” There were several reasons, he said, as to why Bad JAMA and other journals were at least as much to blame as Bad Pharma for a lack of transparency in pharmaceutical research: the constant need of the medical press to make a sensational impact, “the vanity and ambitions of scientists,” and the confusing restrictions of embargos—as well as the fact that, despite the evidence, it was clear that journals do favour “exciting” research. Instead of journals, Professor Senn claimed, trials should be self published either on the web or in some publicly searchable registry, such as the website Clinicaltrials.gov.

And what about the regulatory authorities? On the one hand, Dr Evan Harris, AllTrials campaigner and former LibDem MP, described them as hand in glove with the industry. “Relying on drug regulators is as reliable as relying on the Press Complaints Commission,” he said. He pointed to the failure of the regulatory bodies in establishing the truth about the MMR scandal. “They did absolutely nothing; it was left to a medical journalist to ascertain the truth behind the research paper,” he told the meeting. Yet according to Professor Senn, it’s the regulators, virtually alone, that keep medicine safe. “Regulators may make mistakes, but they do a better job than the journals,” he said. “Would you want to fly to New York with a big reputable airline like BA, which is heavily regulated? Or a plane built by Professor Smith and his colleagues from the local university?”

Even peer review, the process by which research papers are evaluated by independent experts, and which doctors see as a guarantee of high quality research, provoked disagreement. Anonymous peer review, carried out by the vast majority of journals, means that “there is little accountability for the decision to publish the article and no one outside of the journal knows how editorial decisions were made,” said Dr Stephanie Harriman, deputy editor of the open access publisher BioMed Central. Since 2000, the company has pioneered open peer review where the reviewer is named, and his or her review is linked to online publication—increasing transparency and, thereby, accountability. All of this, she said, “means that doctors can make more informed decisions as they are able to look at whether the research was reviewed correctly, and whether they can believe the results.”

With so much to disagree on, speakers and audience members agreed that transparent clinical research is a complex goal, and should be addressed as such. Discussing the future is just the start of the process, pointed out Dr Groves. “Publication bias is not only down to publishers, it is also dependent on people submitting their results including old data, whether it’s in a loft or on a floppy disk or filed away somewhere—so bring out your dead,” she said. “We need to be able to make decisions on all the evidence. That means that observational studies should be regarded as being as important as randomised controlled trials. We know we’ve got to improve and there’s a long way to go. It’s an exciting time,” she said.

Jane Feinmann is a freelance medical journalist and copywriter based in London. She writes about patient safety issues including pharmacovigilance, mental health, ageing, and women’s issues. She organised the MJA meeting discussed in this blog. 

Competing interests: The author has no competing interests to declare. 

  • Deborah Mayo

    I think Senn is right to point to bad JAMA rather than bad pharma. As we saw in the Potti-Nevins Duke case, for example, even the few journals that claim to require data, codes, evidence of reproducibility and so on, do not enforce these requirements, and their desire for sensational stories may result in a conflict with sufficiently careful vetting. Then, when scandal erupts, at least when statistics is involved, too often science writers are ready to pounce on the very methodology that unearths and substantiates the poor error probabilities.

  • freecell0sd

    It is not just pharmaceutical companies that cannot be trusted, but also those promoting psychological and behavioural interventions: http://blogs.plos.org/mindthebrain/2014/05/20/meta-analyses-done-promoters-psychological-treatments-tainted-meta-analyses-done-pharma/

    More raw annonymised data should be made available to allow various analyses and interpretations, so that patients themselves can then choose which approach best reflects their own values and preferences.