When a clinical trial fails, everybody loses: the patients who participated hoping to benefit, the patients who didn’t participate but hoped to get the promising new drug once it hit the market, the researchers who dedicated thousands of hours of their time to it, and of course the trial sponsors, who are usually out many millions of dollars with nothing to show for it. With all that pain, it’s perfectly understandable that the sponsors and researchers often don’t get around to publishing their results.
Understandable, but wrong. As two clinical researchers argue in a commentary in last week’s Science Translational Medicine, the tendency to bury negative clinical trial outcomes makes a bad situation even worse. In an accompanying press release, they explain:
In this situation “scientific information on the efficacy — or lack of efficacy — and safety — or lack of safety — of the investigational agents is not available to the research community, and the opportunity to learn from unsuccessful clinical trials is eliminated,” [Michael] Rogawski says.
For example, Rogawski says that it is assumed that the mouse models used to identify new drugs to treat epilepsy have high predictive value, because every marketed antiepileptic drug has demonstrated activity in the screening models. But “this assumption could be erroneous, because we do not know if there are drugs that were effective in the models but did not exhibit efficacy or had unacceptable side effects in clinical trials and were therefore terminated by their sponsors.”
Besides sandbagging subsequent work, the authors argue that failure to report negative results is unethical. Many patients sign up for a trial and put themselves at risk on the assumption that their sacrifice will do some good, furthering humanity’s understanding of disease. Sponsors who bury negative results are breaching that contract.
Rogawski and his co-author, Howard Federoff, go on to argue that the FDA has the power to change this unfortunate situation. Apparently, the FDA Amendments Act of 2007 may have given the agency the authority to mandate that all clinical trial results be published, regardless of whether the trial met its goals.
It’s always hard to tell how accurate scientists are when they’re trying to interpret legal statutes, but I hope Rogawski and Federoff are right, and that other science regulators are paying close attention. That’s because the “positive bias” in publication isn’t limited to clinical trials; it’s afflicts all of science.
Indeed, people have been moaning for years about the tendency to bury negative results, and the corrosive effect that has on future research. So far, though, efforts to correct it have been entirely voluntary. If regulators actually have been given the authority to force publication, it could be a big step forward. Failed experiments still won’t be much fun, but if we can learn something from them, at least they won’t be complete losses.