Furious Seasons

October 02, 2009

More Than 50 Percent Of Clinical Trials Not Properly Registered

First, I need to apologize for not catching this JAMA study that came out in early September, since it is indeed serious stuff: a study by David Moher of the Ottawa Hospital Research Institute found that many researchers are in violation of a 2005 rule requiring that all clinical trials be registered in a public database, ClinicalTrials.gov. The rule was put in place by the International Committee of Medical Journal Editors and was a precondition for member journals publishing researchers' studies.

In reviewing 323 published clinical trials published in 2008 involving cardiology, rheumatology, and gastroenterology, only 45.5 percent (147 trials) were registered before the end of the trial and had the trial's outcome properly specified. That means that 54.5 percent of published clinical trials were not properly registered: 27.6 percent had no trial registration whatsoever; 13.9 percent were registered after trial completion; and, 12 percent were registered with no outcome measure or an unclear outcome measure.

Stunningly, 31 percent of the trials that were considered adequately registered had a discrepancy between the outcome measure at registration and what outcome measure was actually published. So much for integrity in science.

This is so unacceptable that I don't have words for it. I'm sure there are many journal editors who are furious right now. And many researchers who look stupid.

Not that the media looks much better, since it almost entirely missed Moher's study. The New York Times had a teensy mention of it deep inside the paper and the Ottawa Citizen had an article on the study. That's it.

Lame all the way around.

Posted by Philip Dawdy at October 2, 2009 12:05 AM

del.icio.us Digg it reddit

Comments

Do you call it scientific? Is it the "science" that guides prescriptions?
Is it evidence based medicine?
This is unacceptable to say the least.
I keep on being "anecdotal".

Posted by: Ana at October 2, 2009 03:03 AM

Likely, the editors look stupid too. Journal editors should have checked for inclusion in ClinicalTrials.gov as part of their pre-publication review given that their organization, the International Committee of Medical Journal Editors, issued the rule.

Posted by: Joe at October 2, 2009 07:31 AM

Great Post Philip!!!

The findings are highly disturbing but surprising? Sadly, no.

Keep up the good work!!

Posted by: Angie at October 2, 2009 01:56 PM

This needs to be investigated further. The 'researchers' must have known what the rules were.
Not following the rules, and the glaring outcome jive, there must be rampant crime involved.
We know the pharms are hiring guys like Joe Beiderman to write up phony papers to promote deadly drugs.
Why do we have a DEA to crack down on some drugs, yet the government is doing a darn thing about a problem that's a thousand times worse?
Because it's all about money money money.
The pharms are paying off gangster government scum to let this atrocity continue.
What do we have to do to take down this consortium of corporate and government crime?

Posted by: survivor030406 at October 2, 2009 03:34 PM

"I'm sure there are many journal editors who are furious right now. And many researchers who look stupid."

No you've got it the wrong way round... Trial registration is meant to stop researchers from fiddling their data and not publishing data they don't like. Which is a clever strategy if you're a researcher who wants to get lots of "good" papers but it's bad for science. Journal editors are meant to refuse to publish things that weren't registered.

This shows that researchers are still able to get away with it. Researchers will be happy, journal editors are the stupid ones because they vowed not to publish un-registered trials and it turns out they are still doing so.

Posted by: Neuroskeptic at October 3, 2009 01:01 AM

It's called Due Diligence... Don't assume, it makes an ASS out of U and ME. Yes, I know it sounds trite, but it appears to be true! As somebody mentioned, it would have been simplicity itself for the journals concerned to check the dbase, to make sure that the relevant trial had been properly registered, but whoever was responsible assumed, or turned a blind eye, or just didn't have a procedure in place. Like Philip wrote: LAME. And I'd like to add: UNPROFESSIONAL, and SHODDY.

In the UK, it is a (relatively minor), criminal offence to fail to register a trial (and there's a whole bunch of stuff about advising termination and SAEs, and whatnot, too). Can somebody do a similar study of papers published in UK journals, please!? Because I'll bet that our regulator's dropped the ball on this one, too.

Matt

Posted by: Matthew Holford at October 3, 2009 04:35 PM