Disappointing Cancer Study Results Go Unreported 77
An anonymous reader writes "Science News reports on a new study showing that most cancer drug trial results are never published, probably leaving patients vulnerable to cocktails that have already been shown to be dangerous or useless."
Financial incentive (Score:4, Interesting)
The obvious reason is that it takes time and money to publish study results, neither of which are recouped currently if the study showed negative results.
The obvious fix is to reward pharmaceutical companies financially for publishing all results. Form a subentity within the NIH with the power to purchase study data and results that can be published by the government or a peer-reviewed journal.
Re:Cancer treatment is a product (Score:3, Interesting)
Horse urine can't cure cancer, although it is used in hormone replacement therapy. I suppose some cancer patients driven to early menopause due to radiation and chemotherapy do use it, though...
http://en.wikipedia.org/wiki/Premarin [wikipedia.org]
Re:More complaining and second-guessing (Score:4, Interesting)
As somebody who was previously in the academic medical field, this is not just a problem with cancer drugs. "Positive Outcome Bias" or "publish bias" is a huge problem.
http://www.ama-assn.org/public/peer/7_15_98/jpv71042.htm [ama-assn.org]
A negative study should be just as important as a positive study. If done well... obviously.
Published negative studies dissuade doctors from using certain offlabel treatments. Published negative studies prevent other docs from wasting time and money to discover the same results.
More importantly, many clinical changes are based on meta-studies... which as basically studies which combine all the available data. If negative studies are not published, it throws off these metastudies... and thus bad care occurs.
Re:Negative results (Score:5, Interesting)
I have been told by someone who would know (psychiatrist) that drug trials (maybe just here in Australia) have to be announced before the trial begins otherwise the outcomes can't be published. The idea is that you have to announce a trial before you begin, and so if the outcomes are bad and you don't publish, the bad outcome can be inferred from the lack of publication, even if the specifics remain unknown.
Re:More complaining and second-guessing (Score:1, Interesting)
From the article... there is a registry of all of these trials (from the US, at least). Any group that decides to run a trial is going to check this registry first. If they find that a trial was registered, but no paper was ever released they would probably: a) assume the trial was not a success or b) contact those involved with the trial and ask for more details.
Medical journals are as much for practicing MDs as they are for researchers. Flooding them with failed trials would only dilute the valuable material that MDs rely on to better treat their patients.
I agree it would be nice to have a global registry like this, or even a set of journals dedicated entirely to failed trials.
This is "dark data" (Score:1, Interesting)
Some scientists are interested on getting all data published, even if it didn't result in big news. They call this the "dark data".
http://www.wired.com/science/discoveries/magazine/15-10/st_essay
Plus. (Score:4, Interesting)
There's a wide assumption that the researchers themselves really want to publish all results.
Unfortunately, as in almost the entire field of "science" nowadays, it's not the case.
Researchers themselves have a tendency to hide failures - given that most experiments result in failure, they tend to focus on reporting the ones reporting success.
This use of time simply makes most sense - they don't have the time to report all the failures, and reports of failures not as valuable as reports of success only makes it worse - think about what kind of views your peers will have towards you if most of your publications are negative results.
Sadly, this thinking is parasitic and is very prevalent across all research fields.
Journals are very selective given the limited number of pages they have. If I were a journal, I'd pick reports of success first. It's the evil of centrally-controlled publication, and the mindset that, "if a research is of any good value, it must appear on some journal".
Granted, peer review is a good thing, but there must be a way to give researches credibility without getting published on some journal.
Compounded with big-pharma-sourced funding with very fine strings attached...we have a really screwed up system.
you're dangerously wrong (Score:2, Interesting)
Even if only half of the unreported studies were because of poor results, that's enough to skew things very, very badly.
The basic idea is that you should only use drugs or drug combinations for which there is evidence that they work and are not harmful. If there's nothing published, don't use it!
but by not publishing results they're creating fraudulent overall data, with possibly deadly results. This needs to stop.
There's nothing "fraudulent" about it. Studies often fail for many reasons completely unrelated to the drugs themselves.
Scientific experiments are usually one-sided: a positive result tells you something, a negative result tells you nothing.