Medical journals: an extension of the marketing arm of drug companies?
Everyone knows that medical journals and drug companies are economically interdependent. Drug companies need journals to publish the results of their clinical trials showing the efficacy and safety of new drugs in order to obtain product licences for these drugs. Journals obtain substantial income from drug advertisements, reprints and supplements. "So what?" you may say, "isn't this how it is with all products advertised through the print and broadcast media, so why make a fuss?"
After accepting the 2004 HealthWatch Award Dr Richard Smith, formerly editor of the British Medical Journal, explained to his audience why it is very necessary to make a fuss, reports John Garrow.
The relationship between medical journals and the pharmaceutical industry is far more intimate and sinister than that between advertisers and the general media. It is not only Smith that says so. He cited some important commentators in support of his case: Lancet editor Richard Horton (1) claims, "Journals have evolved into information laundering operations for the pharmaceutical industry". Marcia Angell, former editor of the New England Journal of Medicine, has commented (2), "[The pharmaceutical industry] has moved very far from its original high purpose of discovering and producing useful new drugs. Now primarily a new marketing machine to sell drugs of dubious benefit, this industry uses its wealth and power to co-opt every institution that might stand in its way, including the U.S. Congress, the Food and Drug Administration, academic medical centres, and the medical profession itself." The promotional blurb of a book by Jerry Kassirer (3), another previous editor of the NEJM, reads, "Dr Jerome Kassirer offers an unsettling look at the pervasive payoffs that physicians take from big drug companies and other medical suppliers, arguing that the billion-dollar onslaught of industry money has deflected many physicians' moral compasses and directly impacted the everyday care we receive from the doctors and institutions we trust most."
These are eminent medical editors whose warnings we should heed. If, as Kassirer states, the enormous economic power and influence of drug companies has "deflected many physicians' moral compasses" to the extent that the public is now fed misleading information about the efficacy of drugs, this is something HealthWatch should mind about very much indeed. A few simple figures illustrate the great economic power and profitability of the drug industry. They are now producing relatively few new drugs, but concentrating their efforts more on marketing existing drugs. US companies spent $15.7 billion on drug promotion in the year 2000. That represents about $10,000 on each individual doctor. Since 1995 research staff numbers have been reduced by 2%, while marketing staff have increased by 59%. Researchers comprise just one in five of drug company staff - they are outnumbered two to one by marketing staff.
At present prescription drugs cannot be advertised in the general media (although this rule is being undermined by advertisements on the internet) so the industry needs to persuade doctors to promote and prescribe new drugs in place of older and less profitable ones. It has been shown that endorsement by doctors is more effective in altering prescribing practice than an equal expenditure on company representatives. Independent expert reviewers have found that about half of the advertisements for prescription drugs in medical journals are biased in favour of the drug, giving too great prominence to claims of efficacy, and too little to side effects, and that often the advertisement recommended the use of the drug in a patient group other than that in which it had been tested. It would be possible to peer review all advertisements in medical journals, but this would be very expensive and most editors would rather spend the money on maintaining as high a standard as possible in the research publications. It is not unknown for advertisers to strike a deal with editors, such as favourable editorial mention of a drug in return for placing an expensive advertisement. Some journals carry a section on "product news" which appears to be independent but is in fact "advertorials". And as we heard from the winner of the 2003 HealthWatch Award, Dr Peter Wilmshurst, it may be very difficult to publish reports of an adverse drug effect in major journals, because the makers of the drug will fight vigorously to suppress any such publication.
But perhaps the most important, and certainly the most subtle, way in which the drug industry can influence the opinion of doctors about the efficacy of a drug is by the publication of clinical trials in reputable journals. Systematic reviews of randomised controlled trial are the very best evidence on which Evidence Based Medicine is based. But the systematic reviews are only as reliable as the trials that the meta-analysts have analysed. Of course if trials are technically poor (bad randomisation, weak blinding, inappropriate statistical analysis, etc.) they are given little weight in good systematic reviews, but if the trial is technically perfect we must believe it, mustn't we?
It is unthinkable that reputable pharmaceutical companies and reputable medical journals would collude to publish fraudulent results about a drug trial, but we should bear in mind the pressures that operate when a major drug trial is submitted to a journal for publication. From the journal's viewpoint the financial benefits of publishing the trial are very large (see below). From the drug companies' viewpoint the stakes are even higher. They have already spent many £m in developing the drug, but if a major multicentre trial shows that the drug is relatively ineffective, or has serious side effects, that is a commercial disaster that must be avoided if that is at all possible. Initial drug trials are usually designed and funded by the manufacturer of the drug, the design of the trial is beyond reproach, and almost always the results are favourable to the sponsor. However, when independent researchers study the same drug the results are usually less favourable, and in some cases the drug is withdrawn because it is shown to be ineffective or unsafe. How can this occur?
Dr Smith led us through the methods that can enable companies to get the results they want without falsifying the data:
No need to falsify data: ways in which companies might use real trial results to get the results they want
The new drug can be compared with placebo, or too low a dose of a competitor drug, so the new drug is shown to be "effective", when really it is no better than an alternative treatment.
The new drug can be compared with too high a dose of a competitor drug, so it can be seen to have fewer side effects.
The new drug can be compared with a better (but more expensive) drug in a small trial so the results show "no significant difference" and the new drug appears good value for money.
The trial may have several different end-points, and the report cites those results in which the new drug performed well, but not those in which it performed badly.
The drug may be tested on a heterogeneous group of patients, some of whom did well and others badly. Select a group (eg. men over age 50) who did well and publish those results and forget the rest.
If there is no subgroup that does well do not publish that study at all.
If you have a good study, publish it more than once.
Sponsor multicentre trials, but publish only those centres that show favourable results.
Publish separately different outcome measures from the same trial.
Publish different follow-up periods, eg. results at 3 months, one year, two years...
Publish positive results in major journals and negative or neutral results in minor journals.
Combine results in ways that are favourable.
As an illustration he cited one particular drug about which there were publications describing 84 trials on 11,980 patients. In fact there were only 70 trials involving 8,645 patients, but 17% of the trials had been published more than once, though this was impossible to tell from the published studies. Smith used a Cochrane-type diagram to illustrate the way in which duplicated trials could increase apparent effectiveness. Initially 16 trials (group A) showed that the Number Needed to Treat (NNT) to obtain one favourable result was 9 patients (confidence interval 7-16). Three of the most favourable trials were duplicated (group B): now analysis showed only 4 NNT.
Next group B was duplicated again to give group C, so combining B+C gave 9 trials with 4 NNT, and finally combining all the trials and their selected duplicates there were (apparently) 25 trials giving 5 NNT (CI 4-6), which is a considerable improvement on the initial 9 (CI 7-16). Further examples were given of drugs that had apparently favourable clinical trial evidence (for example Cox-2 inhibitors vs NSAIDs, or HRT to protect against coronary heart disease) but scrutiny by independent experts showed the evidence was flawed. Setting the record straight about the efficacy of drugs is not a task for the faint-hearted: those with an interest in selling, say, Cox-2 inhibitors or HRT will fight fiercely to discredit any evidence that their value has been overstated.
At this stage of the address Dr Smith had convinced us that his title proposition was true: to a very great extent medical journals are an extension of the marketing arm of the drug companies, and someone should do something to correct the situation. But what about those pillars of society - physicians sworn to serve only the interests of their patients? Or the medical press - are they not part of the Fourth Estate, champions of the people, said by Edmund Burke to be more powerful than parliament itself? Why do they condone this scandalous abuse of public trust? As I looked around the audience, many of whom were, or had been (like myself) physicians and medical editors, there were no confident smiles to be seen. A lone representative of the pharmaceutical industry was not looking very happy. How had we got into this mess, and how could we get out?
The first question was easily answered. Reprints of important drug trials, or supplements sponsored by drug firms, are major sources of revenue for journals. Two-thirds of trials in major journals (Archives of Internal Medicine, JAMA, Lancet, New England Journal of Medicine) are funded by the drug industry - for the British Medical Journal it is only one third4. Editors of journals (or their publishers) cannot afford to reject everything that is commercially sponsored. In many fields of medical research (such as obesity, in which I have experience) industry is virtually the only source of funding to employ research registrars, or buy expensive equipment. Unless (like me) they are fortunate to have departments funded by a charity, the great majority of physicians depend upon industry for their research bread and butter, never mind the champagne and canapés that may also be on offer. Politicians call it "partnership with industry" but it is a partnership in which power lies with the commercial sponsor.
What is the solution? Dr Smith offered some answers: A register of trials, so unfavourable trials do not "disappear". Publication of online journals not beholden to commercial sponsorship, such as PLOS Medicine (http://medicine.plosjournals.org); critical review of trial protocols by independent experts; and of course much more public funding of clinically-important trials. Whether his call will be heeded remains to be seen, but it will be certainly receive support from members of HealthWatch.
At the end Dr Smith received a standing ovation, and coped with another 30 minutes of well-informed and pertinent questions. Truly he is a worthy recipient of the 2004 HealthWatch Award.
1. Horton R. New York Review of Books; March 11, 2004.
2. Angell, Marcia. The New York Review of Books; July 15, 2004; 51 (12).
3. Jerome P. Kassirer. On the Take: How Medicine's Complicity with Big Business Can Endanger Your Health. Oxford University Press, October 2004.
4. Egger M, Bartlett C, & Jüni P. Are randomised controlled trials in the BMJ different? BMJ, 2001; 323: 1253.