Okay, not really—but still, unsettling news from yesterday’s NEJM. According to a report from professors at Oregon Health and Science University, Kent State University, and Harvard, negative studies of anti-depressant publications are much less likely to be published in research journals than positive ones.
The authors compared 74 FDA studies for 12 antidepressant agents, which involved a total of 12,564 patients, to those published in medical journals—the goal being to see which studies made it from the insider circuit to the printed page, and how those that did make the jump were altered for publication.
They found that among the FDA-registered studies, “studies viewed by the FDA as having negative or questionable results were . . . either not published (22 studies) or published in a way that . . . conveyed a positive outcome (11 studies).” In other words, bad news didn’t make it to the journals—or it was spun to sounds like good news.
Indeed, according to the study, if a reader were to judge the medications solely based on the published studies, they would think that “94 percent of the trials conducted were positive.” But in fact, “the FDA analysis showed that 51 percent were positive.”
Obviously, medical journals are publishing research that over-sells the effectiveness of anti-depressants. But why? The answer is exactly what you think: a big, fat conflict of interest. Medical journals have a lot to gain by giving designer drugs the benefit of the doubt—even if it’s undeserved.
In a well-known 2006 BMJ study, professors from York University and the University of Medicine and Dentistry in New Jersey, analyzed the messy world of “commercial bias in medical journals.” The risk here is self-explanatory: sometimes the business of medicine encroaches on medical research, especially when it comes to prescription drugs.
But despite the ubiquitous for-profit interests circling around medical journals, the BMJ study found that only nine of 30 peer-reviewed general and internal medical journals “had an explicit policy for dealing with editors’ financial conflicts of interest.” The rest pretty much just wing it—a dangerous strategy when their publications are so economically vulnerable.
Yet medical journals are published by non-profit organizations that,
like any other not-for-profit endeavor, are sometimes strapped for
cash. Companies—particularly drug companies—have an excess of funds.
You do math.
One of the biggest carrots that drug companies can dangle before
medical journals is advertising revenue. A 1999 study in the Western
Journal of Medicine found that most big-league medical publishers see
more than 10 percent of their total revenue from advertisements.
Particularly compelling cases include the journal Clinical Infectious
Diseases (which gets 31.3 percent of its revenue from advertisements),
NEJM (21.3 percent), the Journal of American College of Cardiology
(13.8 percent), the Annals of Internal Medicine (12.9 percent), and
JAMA (10.4 percent). Keeping advertisers happy is essential to these
publications’ bottom lines. And as you might imagine, publishing
studies critical of prescription drugs does not keep pharmaceutical
companies happy.
In fact, drug companies are profoundly sensitive to dissent on the part
of medical journals. In 1992, the Annals of Internal Medicine published
an article that was critical of drug advertisements in journals. The
result was a major drop in drug advertisements in the major medical
publications—with the Annals of Internal Medicine hit hardest. (See char below)
It’s entirely possible that the drug companies held off on advertising
because they wanted to let the hubbub die down. In other words, the
hold-out may have been more about their PR than about punishing the
journals. But that’s exactly the problem: regardless of the motivation,
the financial fate of medical journals is tied to drug company money.
The same holds true for the issue of reprints, which equals big
business for the journals. When a good study is published in a medical
journal, it becomes a cash cow for a good, long while: experts order
copies for citation, anthologies ask for copies, and more importantly,
so do drug companies who want to circulate good news about their
products.
Consider how hard drug companies work to convince the public and the
medical communities about the effectiveness of their drugs. One
prestigious, high-profile study is all that GlaxoSmithKline needs for a
public relations bonanza—the sort of flurry of activity that requires
many, many copies of the relevant study. Ultimately, companies’
obsession with marketing their credibility aligns with journals’
interests in marketing their reprints. It’s a win-win situation for
everyone except the rest of us.
The frustrating thing about all of this is that the journal isn’t
really being insidious—it’s just trying to maintain financial
viability. Non-profit operations are precarious ventures, and often
demand tough choices. As a former editor of BMJ is quoted as saying,
editors often have to choose: “publish a trial that will bring $100,000
of profit or meet the end-of-year budget by firing an editor.” A
decent, honest editor could very well find himself picking the
former—even if the trial in question gets spun in the process.
Clearly, medical journals find themselves in a tricky situation. So
what can be done? Besides the obvious “editors should declare their own
relevant conflict of interests” and “journals should consider not
allowing editors to have any direct financial ties,” the ’06 BMJ study
suggests the following:
- Journals should publish detailed information about their sources
of income, including the amount they get from the sale of reprints,
supplements, and advertising; - The International Committee of Medical Journal Editors (ICMJE) or
a similar organization such as the World Association of Medical Editors
should design a template for journals that would record the wide range
of financial interactions that authors might have with commercial
entities and journals should publish these authors’ statements in full,
either in print or electronically, in conjunction with the articles; - Editors should consider publishing earlier versions of
manuscripts on the Internet together with reviewers’ and editors’
comments, so that readers can see how the tone and content of articles
was changed. This information could help detect the origin of any bias.
These are good ideas, to which I would add a few. First, I’d like to
make another quick point about the utility of the Internet. Today,
small organizations are not as beholden to large ones in order to make
a splash. There is a lot that journals can do online without the help
of drug companies to get their content in more people’s hands: keep
articles free and accessible, create comprehensive trial databases,
grow the online community of experts, etc. There is at least some
potential for the digital revolution to cut out the Big Pharma middle
man when it comes to circulating journal content.
Further, the FDA needs to make a real effort to make its clinical
trials readily available to bridge the information gap between the
trials’ results and what ends up on the printed page. The Internet is
critical here as well. You can imagine an ongoing electronic system
that matches up FDA trials with published trials—in essence, an
automated version of the recent NEJM study.
Of course, such an initiative would require computers—which, according
to one study from this past November, the FDA doesn’t even have. As
Maggie noted in a previous post,
the FDA is on its last legs, under-funded and under-staffed. At the end
of the day, this is no doubt part of the problem: even if it wanted to,
the FDA can’t exercise any meaningful oversight. If we want to make
sure that the information our doctors get about new drugs is true,
regulators need to be given the support they need to do their jobs.
Anticipating a lucrative market for medications, pharmaceutical manufacturers have invested hundreds of millions of dollars in the development of treatments for every degree of potentially disabling medical condition.
One must be objective and decide at what point the benefits of such drugs “truly” outweigh the risks they present. Of course, some of the risks are minor and others, though serious, are tolerable in that they are reversible, short-term or non-life threatening. Many of the risks, however, are extremely serious and are difficult to balance against anything but the most severe symptoms. Clearly, for some people the risks may be acceptable, for others, not.
Drug companies have developed drugs aimed at the widest possible population. That is the most profitable strategy but one that ignores a basic fact in biology, people are different. Bringing a new drug product to market is an expensive business costing tens of millions of dollars. It takes place in a culture of maximum possible sales for maximum possible profit, although the fact is that most drugs don’t work for most people.
Even when a drug has been approved in terms of safety and efficacy, whether it does what the label says it should do, few people realize just how poorly they perform in real life.
Let’s not forget what usually happens when unfettered greed overcomes business fundamentals. Greed and certitude in the power of the dollar have once again clouded judgement.
Niko-
You completely misunderstand how the journal publication process works. The reason for the disparity found in the NEJM article is, for the most part, not bias on the part of the journals; it is bias on the part of investigators. Researchers are much more likely to submit a study to a journal if it is positive than if it is negative. This “publication bias” is a well-known phenomenon and is in no way limited to studies of pharmaceuticals. Drug companies are not the primary villains here although they will tend to facilitate the publication of a positive study and impede the publication of a negative one (they did so much more brazenly in the past but that’s a different scandal).
It is also misleading to compare the overall proportion of positive studies sent to the FDA to the overall proportion sent to journals. You need to stratify the analysis by indication. Many unpublished negative studies are for failed
indications for which no one is any longer interested in using the drugs. Obviously, negative studies are more likely to occur for failed indications than accepted ones.
hello friends I really liked this information, a few days ago I read something similar, I would like to receive updates on this issue, as it is very interesting, thanks!