Researchers found that doctors shown summaries of hypothetical drug tests lost their confidence in the results when a pharmaceutical company had paid for the studies.

Doctors were less willing to prescribe the drug under study, and they even downgraded the very scientific methods they would praise when drugmakers hadn't backed the research.

"I think skepticism is appropriate, but at some point the methods of the research should provide reassurance," said Dr. Aaron Kesselheim of the Harvard Medical School in Boston, who led the work.

Several studies have shown reports in medical journals often cast drugs in a flattering light when the sponsor has money riding on the results. There are also concerns that companies may rig clinical trials in subtle ways to boost the chances that their drug will shine, or spin or hide unfavorable data.

That was the case for GlaxoSmithKline's diabetes pill Avandia. As part of the British drugmaker's $3 billion healthcare fraud settlement earlier this year, the company conceded it had failed to provide safety data to the U.S. Food and Drug Administration and had made unsubstantiated safety claims.

Another controversial drug, Merck & Co's painkiller Vioxx, was recalled in 2004 after it became clear that it had dangerous side effects. Four years prior, researchers backed by Merck had seen a five-fold jump in heart attacks in patients taking Vioxx compared with naproxen, an established painkiller. But they had framed it as a five-fold decrease in heart attacks with the older drug.

One of the ways scientific journals have tried to address this perceived bias is by making sure any potential conflict of interest, such as financial ties to drug companies, are disclosed in research reports.

But it isn't totally clear what effect that has on readers, Kesselheim said.

LITTLE CREDIBILITY

His team came up with a set of imaginary FDA-approved drugs - lampytinib, bondaglutaraz and provasinab - and then created "abstracts" describing clinical trials of each of the drugs as if for a scientific report.

The researchers also varied the funding source and the quality of the research methods, ending up with 27 different abstracts of hypothetical drug tests.

Overall, the 269 doctors surveyed were able to separate the low-quality trials from the better ones. They also had less confidence in drugs studied in low-quality trials and were more reluctant to prescribe them.

Regardless of research quality, however, doctors tended to distrust trials with support from pharmaceutical companies as opposed to no funding or funding from the National Institutes of Health.

For example, they were only half as willing to prescribe a drug when the trial had been sponsored by the industry than when the government had paid for it.

"There was this lack of credibility of industry-funded trials," Kesselheim told Reuters Health, adding that the skepticism could make it harder for new effective medications to gain a foothold.

Matthew Bennett of Pharmaceutical Research and Manufacturers of America, a trade group, said collaboration between industry and academia is critical for the discovery of new drugs.

"Unfortunately, when taken out of context, publicly available information about physician interactions with biopharmaceutical companies does not convey the value of these relationships, which advance science and promote high-quality patient care," he said in an emailed statement to Reuters Health.

TRUST OR NOT

In an editorial published along with the survey in the New England Journal of Medicine, editor-in-chief Dr. Jeffrey Drazen said the journal has taken steps to improve the reporting of clinical trials.

Those steps include requiring that the study is registered in a public database from the outset, publishing study protocols for clinical trials and posting any conflict of interest authors might have, according to the editorial, entitled "Believe the Data."

Those measures aren't enough to satisfy all experts, however.

"We are doing the best we can, but there is still good reason to be skeptical," said Brett Thombs of McGill University in Montreal, Canada, who has studied disclosure in medical journals. "The idea that we can cast aside all doubts and skepticism is naïve."

Thombs said there are several important variables in a clinical trial that can influence its results, such as how faithfully trial participants take their drugs and how the outcomes are assessed, and not all are easy to identify.

"There are numerous factors that are more nuanced than you can get from a few variables," he told Reuters Health. "I don't think we want to throw out a study just because it's funded by a pharmaceutical company, but we are always skeptical."

SOURCE: bit.ly/8H6vG1 New England Journal of Medicine, online September 19, 2012.

• A Permanent Economic Emergency