NEW YORK (Reuters Health) - Scientists who review large sets of drug trials for medical journals often ignore financial conflicts that might warp the evidence, according to a study out Tuesday.
That’s more than just an academic problem, experts say, because the reviews are considered just about the strongest evidence that medical science can muster.
“It influences how physicians make decisions and how guideline panels come up with their guidelines,” said Brett D. Thombs, of McGill University and the Jewish General Hospital in Montreal, whose findings are published in the Journal of the American Medical Association (JAMA).
Thombs’ team found that of 29 reviews, or “meta-analyses,” of earlier drug trials — culled from top journals like JAMA and The Lancet — only two reported who had funded the original trials included in the review.
And none of the reviews mentioned whether the authors reporting on those trials had been paid by drugmakers.
Such financial ties have been linked to research inflating the benefits of new drugs and downplaying the risks, said Thombs.
For instance, according to a 2008 report, only half the trials on antidepressants sent to the U.S. Food and Drug Administration, which approves new drugs, got a positive review by the agency.
By contrast, from the medical literature it appeared that more than 90 percent of the trials favored the drugs, because the majority of those that were unfavorable simply never got published.
To flag potential bias for readers, many medical journals now require authors to declare who funded their study and whether they have any financial conflicts of interest, such as paid consultancies for drugmakers.
Thombs’ team found that more than two-thirds of the original drug trials that ended up being included in the 29 reviews were funded by pharmaceutical companies.
Only 318 of the 509 trials reviewed gave the conflict-of-interest information in the first place.
But most of the time, those financial disclosures got lost in the reviews.
“What we noticed is there is a gap between these two levels,” Michelle Roseman, a graduate student at McGill who also worked on the new findings, told Reuters Health.
When the team contacted the reviewers, the majority admitted they hadn’t even looked at the issue.
“I think it’s a blind spot,” said Dr. Cynthia Mulrow, secretary of the International Committee of Medical Journal Editors, which has published guidelines for how to disclose conflicts of interest in medical journals.
“I do agree that reviewers who summarize evidence should be thinking about conflicts of interest,” she told Reuters Health, “and assessing that as something that could have potentially biased the findings.”
Mulrow is also involved with groups designing guidelines that deal specifically with how to report research reviews.
She said there aren’t any recommendations so far on whether they should report financial conflicts in the original research, but that she and her colleagues would take the new findings into account in future updates.
SOURCE: bit.ly/fil4AO JAMA/Journal of the American Medical Association, online March 8, 2011.