Funding affects researchers' "spin" on results
 

Last Updated: 2002-08-02 12:57:05 -0400 (Reuters Health)

LONDON (Reuters Health) - Authors of journal articles about clinical trials are more likely to draw positive conclusions from their findings if they received funding from a drug company or other for-profit organization, Danish investigators report in the August 3rd issue of the British Medical Journal.

Drs. Lise L. Kjaergard and Bodil Als-Nielsen, from Copenhagen University Hospital, studied all 159 articles on randomized clinical trials published between 1997 and 2001 in the British Medical Journal, a publication that requires authors to declare funding and competing interests.

They examined each report for a link between funding source and the authors' conclusions, specifically the "extent to which overall results favoured the experimental intervention."

Their analysis showed that conclusions of studies were far more likely to show a positive result from the experimental intervention if the study was funded by a for-profit organization.

"The association was quite strong," Kjaergard told Reuters Health. "The interesting thing was that it was only when they were funded by a profit organization alone. If someone else--a nonprofit organization--was involved, somehow, then, the association was not there."

Other sorts of competing interests, such as political, personal or academic interests, did not significantly affect the conclusions.

The study did not determine what was responsible for the link.

"Maybe they [companies] are just a lot better at finding effective treatments. That is one possibility," Kjaergard said. "The other possibility is, of course, bias."

She said a tendency for companies not to publish negative data was also "very likely" to be part of the reason for the association. "But I know that this has changed, and drug companies in the UK are beginning to publish their results regardless of the outcome," she said.

The results emphasize the need for authors to state their competing interests very clearly, she said. "Also, if they have no competing interests, this should also be clear from the report."

The analysis should also serve as a reminder for readers to take conclusions with a grain of salt, she said. "They should do that with all trials. They should always look for bias and it is always a good idea to be critical, irrespective of the funding."

SOURCE: British Medical Journal 2002;325:249-252.

http://www.reutershealth.com/archiv...802elin023.html

What's worse is that the media rarely scrutinizes the information -they'll just repeat whatever the conclusion of the study was. Now that practically all newspapers and TV news shows have a "health" section, you'd think they could hire health correspondants who actually have a background in life sciences who could pick apart the statistics.