How Millions of Experimental Animals Could Be ‘Missing In Action’ Due To Publication Bias

A study has found that out of over 5500 experimental animals used at one university in the Netherlands over a two-year period, only around a quarter were ever mentioned in a scientific publication. The authors believe that this may be a trend in institutions around the world.

Researchers reviewed study protocols submitted to the animal ethics committee by three University Medical Centre Utrecht (UMCU) departments in 2008 and 2009. They then searched for the resulting papers. They found that of the 5590 experimental animals used, just 1471 were mentioned in scientific publications. Small animals like rodents and rabbits were the most likely to ‘disappear’. Only 23% of them appeared in publications, despite making up 90% of animals used.

So why are these animals not accounted for? According to the researchers’ survey, the most common reasons were that studies were not published because they either didn’t reach statistical significance, were part of a pilot study, or encountered technical issues with the animal models.

Allowing the results of a study to influence whether or not not those results are published is called publication bias, and is a major problem in scientific research. Just because a study produces negative results, this doesn’t mean that those results are not valuable to other researchers, who may otherwise waste time and resources repeating experiments that have already failed. This could be argued to be even more true for animal studies, in which the lives of animals may have been sacrificed to produce results that essentially go to waste.

Unlike in human trials, which are much more transparent when it comes to tracking participants, such details in animal trials are usually listed in applications for ethical approval, which often remain confidential. To address this problem, the authors of this study urged researchers to register protocols involving animals in an accessible online database prior to carrying out their experiments. To facilitate this, they have developed a platform dedicated to animal study protocol registration: www.preclinicaltrials.eu. However, the Dutch government has yet to make the use of such platforms compulsory.