What researchers don’t want you to know about their clinical trials


This article was originally on a blog post platform and may be missing photos, graphics or links. See About archive blog posts.

Clinical trials can tell us whether new drugs or medical procedures work better for patients than the old ways of doing things. They can also tell us whether a new treatment is dangerous. But medical researchers routinely emphasize the upside of the intervention they studied while glossing over its risks for patients, according to a new report.

“The reporting of harm is as important as the reporting of efficacy in publications of clinical trials,” a group of French researchers wrote in a study being published in Tuesday’s edition of Archives of Internal Medicine. “However, harm is frequently insufficiently reported.”


How insufficiently?

The researchers examined 133 studies published throughout 2006 in five of the most influential medical journals – New England Journal of Medicine, Lancet, Journal of the American Medical Assn., British Medical Journal and Annals of Internal Medicine. Of those studies, 83% assessed new drugs, and 55% were funded in whole or in part by for-profit companies.

A total of 15 studies, or 11%, made no mention of any adverse events whatsoever, the researchers found. An additional six studies mentioned adverse events but didn’t provide any numerical data about them.

Even when articles did include hard numbers, they were often presented in a distorted way. For instance, 36 studies didn’t distinguish between minor and severe side effects. Seventeen studies reported only the most common side effects, and an additional 16 included only the most severe. Five of the studies restricted themselves to the side effects that were linked to the drug or treatment being studied in a statistically significant way. (Such side effects are rare, since clinical trials aren’t designed to detect them.)

The numbers can also be distorted by focusing on the number of study participants who had an adverse reaction instead of giving the total number of adverse reactions. Of the 90 studies that included numerical data on side effects, 31 did this, according to the French researchers.

Only 17 articles described adverse events that prompted volunteers to withdraw from studies; 63 gave no data on such withdrawals, the researchers found.

The Consolidated Standards of Reporting Trials (a.k.a. “CONSORT”) was amended in 2001 to emphasize the necessity of reporting “all important adverse events or side effects.” But clearly, the French team concluded, “the reporting of harm remains inadequate.”


Dr. John Ioannidis of the University of Ioannina School of Medicine in Greece, was one of the first researchers to study the way safety data are conveyed in medical journals. In an editorial accompanying the French study, he attributed some of the under-reporting to companies intent on “silencing the evidence” that their products could be harmful to patients.

Ioannidis singled out the cases of Vioxx, a painkiller that was withdrawn from the market because it doubled the risk of heart attack and stroke, and Neurontin, an epilepsy drug that the Food and Drug Administration linked to an 80% increase in suicidal thoughts and behavior.

“In these cases,” he wrote, “marketing needs prevail over scientific accuracy and clinical prudence.”

-- Karen Kaplan