« Congo cracks down on illegal logging | Main | On Nature News »

Bookmark in Connotea

Cancer studies sometimes conceal conflicts - May 11, 2009

Clinical cancer research is often conducted by scientists with conflicts of interest, such as ties to the company making a drug tested in a study. And studies conducted by conflicted researchers are more likely to report positive findings, researchers reported yesterday.

The findings come from a study led by Reshma Jagsi of the University of Michigan in Ann Arbor, who published her results in the journal Cancer.

Jagsi and her colleagues studied 1,534 cancer research reports published in eight top journals in 2006. Twenty-nine percent of the studies appeared to have a conflict of interest. However, only 17 percent disclosed a conflict of interest. And randomized clinical trials that measured a treatment's impact on patient survival were more likely to report positive results if a conflicted researcher was involved with the study, Jagsi's team found.

The New England Journal of Medicine had the highest percentage of conflicts, which were present in 61 percent of its 31 oncology studies. Cancer had the lowest ratio; 16 percent of its 602 studies reported conflicts.

Conflicts of interest in research have long been a controversial subject, with critics arguing that journals and researchers don't do enough to keep conflicts from affecting research. Conflicts can bias study designs and data interpretation, they argue.

Jagsi said in a press release that more public funding of research might be warranted in light of her findings. "It has been very hard to secure research funding, especially in recent years, so it's been only natural for researchers to turn to industry. If we wish to minimize the potential for bias, we need to increase other sources of support," she said.

Post a comment

Comments will be reviewed by the blog editors before being published, mainly to ensure that spam and irrelevant material (such as product advertisements) are not published . Please keep your comment brief. Excessively long or offensively phrased entries will be edited.

We strongly encourage you to use your real, full name. E-mail addresses are required in case we need to discuss your comment with you directly. We won't publish your e-mail address unless you request it.

Please enter the numbers you see below - this helps us to cut down on spam. Note that attempting to post within 30 seconds of hitting ‘preview’ or ‘post’ can cause the system to think you are spamming the site. If you are having trouble with this system, you can instead e-mail a comment to 'thegreatbeyond at nature.com'.

please enter code

TrackBack

TrackBack URL for this entry:
http://blogs.nature.com/cgi-bin/mt/mt-tb.cgi/8209