Reviews of drug trials underreport conflicts of interest

2

Do you follow all your doctor's instructions? Should doctors believe all the research from published drug trials? Evidence-based medicine aims to help doctors know what to believe by using tools like standardized, practice guidelines; but even these can sometimes be based on flimsy scientific support. Scientific support is graded from strong evidence like rigorous, systematic reviews to lower priority, expert opinions.

A meta-analysis uses statistics to summarize results from multiple studies, and is one of the most cited study designs and often prioritized in practice guidelines. Since meta-analyses are so frequently and highly regarded, one might think that they would be held to the highest standard of reporting policy- but they aren't.

Funding information and conflicts of interests in randomized controlled pharmaceutical trials are rarely reported when the data is combined into a meta-analysis, according to an article in this week's Journal of the American Medical Association.

Dr. Brett Thombs is an assistant professor at McGill University and one of the authors of the review. In an e-mail to me, he described how industry-sponsored studies, particularly those that are published, are more likely to report results that support the efficacy of the sponsor's drug compared to independent studies.

"Since industry sponsorship is known to be a potentially biasing factor, we wanted to see how often it is reported by meta-analysts and whether they consider it as a potentially biasing factor. We found that meta-analysts almost never report this information and do not consider it in the context of potentially biasing factors," he wrote.

They reviewed meta-analyses from the top impact general medicine journals, as well as the top journals in the 5 specialty medicine areas with the highest drug sales in 2008, and the Cochrane Database of Systematic Reviews. Overall, they reviewed 29 meta-analyses, which included 509 randomized controlled trials (RCT).

They found that most meta-analysts are not even recording whether or not the studies they review are industry-funded versus independent studies or whether authors of these studies have important financial ties to industry. Only 2 meta-analyses reported RCT funding sources and 0 reported author-industry ties or employment by the pharmaceutical industry.

Critics of this review might point out the relatively small number of included analyses or liken reporting conflicts of interest to splitting hairs in high quality meta-analyses, but I would argue that when it comes to standardizing practice guidelines, there is no such thing as too much transparency.

Rebecca Riley was diagnosed with ADHD and bipolar disorder by age 3, and ultimately died from an overdose of multiple psychiatric drugs by age 4. Her doctor had been prescribing Depakote, Seroquel and clonidine, and was brought to trial. The defense team argued that the psychiatrist acted "in full accord with acceptable medical standards" and the psychiatrist was ultimately cleared of criminal charges and continues to practice at Tufts Medical Center.

Cases like hers make me think that if meta-analyses are being prioritized for acceptable medical standards or guidelines, perhaps it makes a difference if they keep track of and report conflicts of interests or ties to the drug industry.

Posted By: Winston Chung MD (Email) | March 08 2011 at 01:01 PM

Reviews of drug trials underreport conflicts of interest http://www.sfgate.com/cgi-bin/blogs/wchung/detail?entry_id=84453 del.icio.us Digg Technorati Reddit LinkedIn Facebook Slashdot Fark Newsvine Google Bookmarks Twitter Share

| Email