This post by David Healy and Peter Selley does not question the good intentions of anyone mentioned. But sometimes good intentions can land us in a mess.
In 2001, Study 329, a clinical trial comparing paroxetine to imipramine and placebo was published; it claimed paroxetine was effective and safe (1). Jon Jureidini and other clinicians drew attention to problems with the publication (2).
Investigative journalists working for the BBC, with clinical input, also became interested in Study 329 (3).
In 2002, GlaxoSmithKline (GSK) submitted Study 329 and two other trials to FDA seeking a license to treat depressed children. They made it clear paroxetine did not work in these trials. In an approvable letter, FDA responded agreeing these studies were negative and agreeing not to mention this in the label of the drug.
In 2004, GSK were charged with fraud by the Attorney General of New York on the basis of a 1998 document BBC journalists had sourced. The document indicated a company intent to pick out the good bits of the data and publish them, primarily to claim paroxetine was effective in Study 329 despite an internal company assessment stating it was not effective.
GSK resolved their New York legal case with a payment and an offer to make the data from their studies publicly available. The company later faced a Department of Justice action, in which Study 329 featured, which was resolved with a record breaking monetary payment.
A number of articles around this time made it clear that company studies were routinely written by medical writers (4). Medical writing (ghostwriting) raised a number of concerns. One was that a medical writer might not represent the study outcomes as a clinician might. A second centred on a failure to acknowledge the input of medical writers.
Two decades ago, BMJ editors commonly attended medical writer meetings. There was an understanding that medical writers produced articles that met quality metrics for publications more reliably than academics and turned around a product more expeditiously than academics. There was a view that writers also tempered the input of company marketing departments. Medical writing continues to this day. The contribution of medical writers is more often acknowledged at the end of published articles. The editors of leading journals and medical writing companies still liaise closely on medical publication matters.
A legal case established that Study 329 had been ghostwritten by Sally Laden. Between medical writing and a mismatch between the data and the publication, there was a strong case to get the publication retracted. Jon Jureidini and Leemon McHenry wrote to the journal, the authors, and their institutions inviting a retraction.
The article remains unretracted. A complicating factor is that over 50% of the studies of antidepressants, in all age groups, submitted to FDA around this time were negative (5,6,7). Many were published as positive, particularly in adolescent trials (7). Another factor is that it is not obvious that there was any research misconduct by the clinicians involved. Study 329 was run in the very best institutions by leading clinicians and to a high standard.
The primary problems lay with GSK’s handling of the study. GSK’s handling, however, was standard company modus operandi and the company could point to the fact that none of the clinicians or their institutions had called for a retraction.
In 2007, to coincide with a further BBC investigation of ghostwriting and the selective reporting of Study 329 results, Fiona Godlee, the editor of the BMJ, approached Jureidini and McHenry to write an article on their failure to secure a retraction of the Study 329 paper. The unretracted paper was being cited in guidelines as evidence for the effectiveness and safety of SSRIs, including paroxetine, and was therefore shaping the treatment of depressed adolescents.
BMJ sent the Jureidini and McHenry article to Richard Smith and Elizabeth Wager to review. They did not support publication, citing historical practices that had now changed, uncertainties about what was being claimed, and legal risks. Jureidini and McHenry elicited BBC support and provided documentary support for all the points they made. To no avail.
Much of the McHenry and Jureidini article was published in 2009 in Accountability in Research without legal consequence (8, 9). In 2010, BMJ commissioned an article overlapping the Jureidini and McHenry article, written by Melanie Newman, listed as linked to the Bureau of Investigative Journalism (10). Newman’s article balanced the original claims with alternate interpretations like Elizabeth Wager’s view that these practices were historical.
In 2006, Healy submitted an article to BMJ on the inappropriate handling of significant adverse events in company adult antidepressant trials. The reviewers, Stephan Evans and Simon Wessely, supported publication. After returning the galley proofs of the article, Healy received an email saying BMJ’s lawyers had stopped publication. It was published a year later as ‘Did regulators fail over selective serotonin reuptake inhibitors’. Only one word had changed – companies became regulators (11).
In 2014, Jureidini, Healy and colleagues submitted a first Restoration of Invisible or Abandoned Trials (RIAT) article to BMJ, restoring Study 329 (12). Over a year later it had not been published. Among other things a legal review, it seems, intimated that by virtue of their prior engagement in these issues Jureidini and Healy were conflicted. It would be better if an entirely independent group analyzed the data. The roadblock to publication lifted, after it was noted that a senior member of the BMJ editorial team had a more striking conflict of interest than Jureidini or Healy (3).
See How to Become a Ghostwriter for Beginners
BMJ and Investigative Journalism
In the 1850s the Lancet all but created investigative journalism publishing features on the contents and health consequences of proprietary foods and medicines. This led in 1875 to the first Food and Drugs Act. By 1900, investigative journalism had become a distinct branch of journalism, embracing the British Medical Journal’s Secret Remedies series in 1910 and 1912 (13).
The BMJ recently set up an explicitly designated investigative journalism unit, whose site lists a string of publications dating back over a decade before the unit was set up (14). Some of the contributions, like the Thacker article on Brook Jackson and Ventavia, are striking (15).
BMJ investigations around 1900 were embedded in obviously medical journals. In addition, adding the word Investigative to Journalism suggests a publication that takes a position, analogous to the need for a doctor to take a position when faced with a possible adverse event on treatment – there is no escaping a decision to increase or reduce the dose of treatment. Is the new BMJ Unit investigative in this sense or does it serve another function?
In December 2022, Peter Selley and David Healy contacted the Investigative Journalism Unit, calling their attention to GSK and Pfizer trials of maternal vaccines for RSV infections. The GSK trial was suspended early in 2022. GSK did not respond to our request for information. The MHRA responded after a fashion, indicating there had been an issue with premature births. In fact, the trial also had an excess of neonatal deaths. GSK had informed investors of this in early 2022 – investors have rights to know about adverse events that clinicians and patients do not have.
A just published Phase 2 Pfizer study of a vaccine, similar enough for GSK and Pfizer to sue each other for breach of patent, pointed to a somewhat concealed, comparable prematurity problem. Subsequent Phase 3 Pfizer data, trumpeted as proving efficacy and safety, are open to question.
Selley and Healy hoped BMJ’s Investigative Team could find out more than they already had. The BMJ unit expressed initial interest, were then slow to engage, and finally made contact through Hristio Boytchev.
An article appeared on May 10th (16). This was short on key details, not mentioning that the consent form for the Pfizer trial had told women their babies could not be harmed as they, rather than their babies, were getting the vaccine. Selley and Healy were told that a Consent article would follow but BMJ lawyers were having difficulties – perhaps women at some sites had been informed there could be hazards. The consent article had still not appeared 3 months later.
More to the point, the article under Hristio’s name only, apart from some sound bites from experts, essentially contained nothing that had not been given to him by Selley and Healy, with no acknowledgement of this. Rapid responses by Selley and Healy to the article, bringing out extra points for consideration, were not published.
The medical writers who had done the work vanished.
A case can be made that in areas where pharmaceutical companies might be sensitive about their products, an article by a journalist rather than a doctor poses less risk to BMJ. There is less of a causal heft to it. The journalist can omit details and elicit a broad base of comments from other experts, as Newman did a decade earlier, to dilute the impact of points being made.
This may be good journalism, but it suggests BMJ may no longer be a journal for members of the British Medical Association. The mission statement of its current leadership team is:
Combining specific industry knowledge with a broad range of experience, our leadership team are dedicated to improving patient outcomes.
At BMJ, our leadership team is represented by the heads of our central business functions and each of our international hubs. The team works closely to make sure that, as a company, we deliver on our strategic goals.
This reads like a mission statement from a publishing corporation, whose ethos aligns more with that of pharmaceutical or life sciences corporations than that of a professional body.
The concern for patient outcomes may be genuinely held, and ensuring the survival of the BMJ to continue tackling issues may help with this, but does managing the risks to the corporation now trump a concern for patient outcomes? Managing risks to a corporation may paradoxically make a pharmaceutical company embrace of medical writing easier to understand and make critiques of company practices by BMJ and other ‘medical’ journals sound hollow.
It also condemns BMJ to finding and throwing rotten apples (individual instances of bad practice) out of a barrel. Disposing of rotten apples risks generating a false sense of security for patients, and those working within what was once healthcare but are now health services, if the problem is a rotten barrel.
A 1991 pharmaceutical company publication in BMJ created what in this case is the possibly pertinent rotten barrel (17). This company article took issue with a clinical article, which, using the appropriate scientific methods to determine cause and effect, established fluoxetine can cause suicide (18). The company claim was that randomized clinical trials (RCTs) are the scientific way to determine cause and effect. They are not. RCTs speak to average effects. In the company article, moreover, there was a significant excess of suicidal events on average on active treatment – which BMJ ignored.
This company publication effectively created Evidence Based Medicine. In publishing it, BMJ gave medical journals and their lawyers difficulties with the word ‘cause’ and with medical determinations of cause and effect tied to specific cases. Medical practice and reports of adverse events conform with scientific norms in a way that company trials do not. Case reports, which embody our best and most scientific data on adverse events, vanished from leading medical journals. Even letters raising cause and effect issues, like those from doctors alerting colleagues to the birth defects caused by thalidomide, are not published today.
A publishing business based on sales of company trial reprints, and articles by journalists, replaced medical journals.
While corporations and professions can work together, their interests can conflict. Nowhere is this more apparent than when a patient has an adverse event on a treatment. Professionals are defined by an obligation to put the interests of the person seeking their help before those of corporations, whether State, Church, or Company. Facing a possible adverse event, doctors have no option but to make a causal determination and act accordingly. If the outcome of their action supports their judgement about a link to treatment, they should inform clinical colleagues and the natural place to do so would be in the journals of their professional associations. Is this still possible?
References
Keller MB, Ryan ND, Strober M et al. Efficacy of paroxetine in the treatment of adolescent major depression: a randomized, controlled trial. J Am Acad Child Adolesc Psychiatry 2001;40: 762-72.
Jureidini J, Tonkin A, Paroxetine in major depression (letter to editor), J Am Acad Child and Adolesc Psychiatry 2003, 42: 514.
Healy D, LeNoury J, Wood J. Children of the Cure. Samizdat Healthwriters’ Cooperative, Toronto 2020.
Healy D, Cattell D. The Interface between authorship, industry and science in the domain of therapeutics. Brit J Psychiatry 2003, 182, 22-27.
Turner EH, Matthews AM, Linardatos E, et al. Selective publication of antidepressant trials and its influence on apparent efficacy. N Engl J Med. 2008, 358:252–60.
Laughren T. The scientific and ethical basis for placebo-controlled trials in depression and schizophrenia: an FDA Perspective. European Psychiatry 2001,15, 418-423.
Healy D, Le Noury J, Jureidini J. Pediatric antidepressants; benefits and risks. Int J Risk & Safety in Medicine 2019; 30, 1-7.
McHenry LB, Jureidini J. Industry sponsored ghostwriting in clinical trial reporting: a case study, Accountability in Research 2008, 15: 152–167.
Jureidini J, McHenry L. The Illusion of Evidence Based Medicine. Wakefield Press, Adelaide, 2020.
Newman M, The rules of retraction, BMJ 2010, 341:c6985.
Healy D. Did regulators fail over selective serotonin reuptake inhibitors. BMJ 2006, 333, 92-95.
Le Noury J, Nardo J, Healy D, Jureidini J, et al. Restoring Study 329: A randomised, controlled trial of the efficacy and harms of paroxetine and imipramine in the treatment of adolescent major depression BMJ 2015; 351: h4320
Healy D. Shipwreck of the Singular. Samizdat Healthwriters’ Cooperative, Toronto 2021.
The BMJ Investigations Unit. https://www.bmj.com/investigations
Thacker P. Covid-19: Researcher blows the whistle on data integrity issues in Pfizer’s vaccine trial. BMJ 2021; 375: n2635
Boytchev H. Maternal RSV vaccine: Further analysis is urged on preterm births. BMJ 2023; 381 p1021
Beasley CM, Dornseif BE, Bosomworth JC et al. Fluoxetine and suicide: a meta-analysis of controlled trials of treatment for depression. BMJ 1991, 303: 685–92.
Teicher MH, Glod C, Cole JO. Emergence of intense suicidal preoccupation during fluoxetine treatment. Am J Psychiatry 1990, 147, 207–10.
The post Are Journalists the Answer to Clinical Medicines Cause and Effect Problem? first appeared on Dr. David Healy.