A new US study led by a researcher who used to review drug studies at the US Food and Drug Administration (FDA) suggests that selective reporting of the results of antidepressant drug trials in journals exaggerates their effectiveness. This could mislead doctors and patients into thinking antidepressants are more effective than they really are said the researchers.

The study is published in the 17th January online issue of the New England Journal of Medicine (NEJM) and is the work of Dr Erick Turner and colleagues. Turner is assistant professor of psychiatry, physiology and phamacology at Oregon Health & Science University (OHSU) and Medical Director of the Portland Veterans Affairs Medical Center’s Mood Disorders Program.

For three years, Turner was a clinical reviewer of psychotropic drugs at the FDA. He has also been a clinical researcher in the intramural program of the National Institutes of Health’s National Institute of Mental Health and has published around 35 papers in peer reviewed journals.

Turner and colleagues looked at FDA reviews on trials of 12 widely prescribed antidepressant drugs that were approved between 1981 and 2004, involving a total number of 12,564 participants.

They also systematically searched the literature to find out whether the studies reviewed by the FDA had been published in medical journals. If they found a published study, they compared it to the version reviewed by the FDA.

They also did a separate meta-analysis across all the studies for each drug to find out its “effect size”, giving them a quantitative way to compare claims made about a drug in the literature with that reported in the FDA version of the trial reports.

They found a significant link between the results of a study and whether and how they were reported in medical journals. They also found significant differences in effect sizes for the drugs.

Specifically, the results showed that:

  • 31 per cent of 74 FDA-registered studies (accounting for 3,449 participants) were not published.
  • 37 of the studies seen by the FDA as having a positive outcome were published, and 1 study viewed as positive was not.
  • Studies viewed by the FDA as having a negative outcome were, except in 3 cases, either not published (22 studies), or published in such a way that the outcome was presented as positive (11 studies). These 33 studies accounted for a total of 5,212 participants.
  • Collectively, according to the published literature, it appeared that 94 per cent of the trials conducted were positive.
  • This contrasted sharply with the FDA view, which showed that only 51 per cent of the trials were positive.
  • The meta-analysis revealed that the increase in effect size from FDA to journal data sets ranged from 11 to 69 per cent for individual drugs and came to 32 per cent across all 12 drugs.

Turner and colleagues concluded that:

“We cannot determine whether the bias observed resulted from a failure to submit manuscripts on the part of authors and sponsors, from decisions by journal editors and reviewers not to publish, or both.”

On the consequences of selective reporting, they added that:

“Selective reporting of clinical trial results may have adverse consequences for researchers, study participants, health care professionals, and patients.”

In a separate statement, Turner said that:

“Selective publication can lead doctors and patients to believe drugs are more effective than they really are, which can influence prescribing decisions.”

He was also careful to point out that just because there is over-reporting of positive results or there is a large number of negative studies, it does not mean that a particular drug is ineffective.

He and his team also worked out from the meta-analysis, that when pooled, the results on each drug were still better than treatment with a placebo.

The claim being made in this study is not that the drugs are ineffective but that the view that is presented in the medical literature makes them out to be more effective than they really are (based on the reasonable assumption that the FDA-registered versions are closer to the “truth”).

Turner said it was important that doctors and patients “have access to evidence that is complete and unbiased when they are weighing the risks and benefits of treatment”.

Some drug companies have criticized the results of this study, saying that it did not take into account papers that reviewed more than one drug, and other ways in which doctors can get hold of trial results, whether positive or negative.

Some said they publish their trial results on their websites, and also that some studies that aren’t published in the journals are registered with online trial registries, for example www.clinicaltrials.gov.

However, many studies are not being registered or reported, said Kay Dickersin, the director of the Center for Clinical Trials at the Johns Hopkins Bloomberg School of Public Health, speaking in the Wall Street Journal today:

“We need something more meaningful,” she told the paper, “The average person has no idea that www.clinicaltrials.gov is not comprehensive.”

“Selective Publication of Antidepressant Trials and Its Influence on Apparent Efficacy.”
Turner, Erick H., Matthews, Annette M., Linardatos, Eftihia, Tell, Robert A., Rosenthal, Robert
N Engl J Med 2008 358: 252-260.
Volume 358:252-260, January 17, 2008, Number 3.

Click here for Abstract.

Sources: NEJM journal abstract and press release, Wall Street Journal.

Written by: Catharine Paddock