Drug Companies Routinely Bury Studies Showing Their Drugs Don't Work

Wednesday, August 20, 2008 by Mike Adams, the Health Ranger Editor of NaturalNews.com (See all articles...) Tags: clinical trials, health news, Natural News

(NaturalNews) More than 30 percent of studies conducted on antidepressant drugs go unpublished, apparently because they fail to show that the drug works as advertised, according to a new study published in the New England Journal of Medicine.

In addition, published studies often interpret the data as more favorable toward the drug than it actually is.

"Not only were positive results more likely to be published, but studies that were not positive, in our opinion, were often published in a way that conveyed a positive outcome," the researchers wrote.

Although the FDA requires that pharmaceutical companies provide the results of all studies conducted on a new drug before it receives approval, there is no requirement that those studies be published where doctors, health researchers and the public can read them. A team of researchers from institutions including the Oregon Health and Science University, the Portland Veterans Affairs Medical Center, Kent State University and the University of California at Riverside compared the results of studies reported in the FDA database with whether and how those studies were reported in published literature.

"It tells you where they placed their bets before they saw the data," lead researcher Erick Turner said.

The researchers examined 74 studies conducted on 12 antidepressants between 1981 and 2004, involving a total of 12,564 patients. Only 38 of these studies (51 percent) produced favorable results for the drugs being tested. All but one of those studies - or 94 percent - were published.

In contrast, 22 of the 36 studies that failed to show favorable results (61 percent) went unpublished. Of the 14 studies published, only three interpreted the data unfavorably. The other 11 studies, in contrast with FDA interpretations of the same data, cast the drugs as more effective than they actually were.

For example, of the three studies conducted on Glaxo SmithKline's Wellbutrin, only the one showing positive results was published. Of five studies conducted on Pfizer's Zoloft, the one showing positive results was published; the one with questionable results was spun as if the drug were effective, and the three with unfavorable results were never published.

According to the researchers, the publication bias inflated the reported effectiveness of the anti-depressant drugs by an average of 32 percent. The actual amount varied from a low of 11 percent for Glaxo SmithKline's Paxil to a high of 69 percent for Bristol-Myers-Squibb's Serzone.

Turner said that this exaggeration of a drug's effectiveness can lead doctors to prescribe it in borderline cases where it might not be appropriate.

"Selective publication can lead doctors to make inappropriate prescribing decisions that may not be in the best interest of their patients and, thus, the public health," the researchers wrote.

The researchers said that the publication bias did not necessarily arise from drug companies actively concealing data. One explanation might be that researchers find negative results uninteresting, and do not go to the effort of writing them up.

"There's an expectation that if you get a positive result, that's what you're supposed to do, and if you get a negative result you have failed," Turner said. "The first impulse is to say, 'I was wrong. Maybe I should move on to something more interesting.'"

The bias could just as easily come from medical journals being more interested in studies with positive results, Turner said, and not from the researchers at all.

Newsweek columnist Sharon Begley expressed skepticism with this view.

"I think he's being too kind," Begley said. "With the proliferation of journals, virtually anything can get published somewhere - maybe not in the New England Journal of Medicine, but in some third-tier rag, which would still count as 'published.'"

Regardless of the reasons behind it, it is widely acknowledged that publication bias has far-reaching ramifications. It skews the results of meta-analyses, in which researchers compile the results of many published studies to get a better picture of how effective a drug or other therapy is. In extreme cases, it can create a totally false impression: Because studies are typically conducted within a 5 percent confidence interval, it is expected that one in 20 of those studies will show a false result simply due to random variation. If only that study is published and the 19 studies showing no results are never seen, this conveys the opposite impression of what the research actually found.

While the current study is the most thorough to date examining publication bias in the drug industry, health professionals have long known that this bias is a problem. In 2004, the state of New York sued Glaxo SmithKline for intentionally concealing the results of studies on Paxil in order to make it appear that the drug was more effective on children and adolescents. The company also sought to conceal evidence of dangerous side effects, the lawsuit alleged, such as an increased incidence of suicidal tendencies.

While that lawsuit was eventually settled, at least 12 jury trials are scheduled for 2008 that accuse Glaxo SmithKline of hiding suicidal side effects in youths.

As part of the 2004 settlement, Glaxo SmithKline agreed to post the results of all its drug trials on the Internet, and other major companies followed suit. In response to the New England Journal of Medicine study, spokespeople for several drug companies pointed to this practice as evidence that the system has been reformed.

But Turner says that posting results online is not the same as publishing them in the medical literature. He doesn't know a single doctor, he said, who regularly consults the companies' online databases.

"We are taught that the medical literature is the Holy Grail," he said.