Quotes of the Day

Peer-reviewed journals wield enormous power in the pharmaceutical industry. A study published in one of the more prestigious — the New England Journal of Medicine (N.E.J.M.), say, or the Journal of the American Medical Association (J.A.M.A.)--can make or break a new drug. But the journals are far from perfect. One big problem, says Dr. Drummond Rennie, a J.A.M.A. editor, is that "while journals are very good at evaluating the significance of studies sent to them, what they don't do well is evaluate what's not there."

And unfortunately, when it comes to data on clinical drug trials, there is a lot that's not there. Why? First of all, researchers who receive industry funding don't necessarily get to decide whether the results are submitted for publication. Many researchers also practice a form of self-censorship in an effort to please their sponsors or in mistakenly believing that journals tend to favor positive results. Others will cherry-pick a narrow slice of data for publication while consigning the rest to the file drawer. Whatever the reason, the result is a bias against negative or inconclusive data that distorts the medical literature and ultimately the practice of medicine.

And that's why there is a movement afoot to require researchers to register all clinical trials when they start. This isn't a new idea, but it's been given new urgency by, among other cases, the example of GlaxoSmithKline (GSK) and its antidepressant Paxil. While data that showed Paxil could help depressed adolescents were published in the Journal of the American Academy of Child & Adolescent Psychiatry, the results of another trial, which raised concerns about children having suicidal thoughts while on Paxil, were not submitted for publication. (GSK points out, however, that the negative results were presented at a major medical meeting.)

A registry won't mean that every failed trial will be published, but it will make it clear that those trials took place. The International Committee of Medical Journals, which includes J.A.M.A. and N.E.J.M., is drafting plans for a registry system, and the A.M.A. has a proposal on the table. Doctors met last week with Democratic lawmakers who are considering legislation that would require disclosure of results. Meanwhile, some drugmakers, among them GSK and Merck, have indicated that they are more inclined to back the trend than to buck it.

The prospect of having data from all clinical trials readily available is exciting, says Dr. Cary Gross of Yale: "This has the potential to dramatically change how medicine is practiced." But just as compelling is the ethical argument. "When people enroll in a study, they are told that this is going to benefit science," says Gross. "For trials results to then be shelved is a real betrayal of people who have trusted researchers with their lives." The pharmaceutical industry would be well advised to see it that way too.

• David Bjerklie