Animal research has a publication problem. About half of all animal experiments in academic labs, including those testing for cancer and heart drugs, are never published in scientific journals, and those that are have been notoriously hard to replicate. That’s part of the reason that most drugs that work in animals don’t work in people—only 11% of oncology compounds that show promise in mice are ever approved for humans—despite billions of dollars spent by pharmaceutical and biotech companies. Meanwhile, academic labs waste money, mice, and other resources on experiments that, unbeknownst to them, have already been done but were never reported.
In response to similar concerns about human studies, the U.S Food and Drug Administration (FDA) in 2007 mandated that researchers conducting human clinical trials preregister the details in an online database like ClinicalTrials.gov. Now, some scientists are wondering whether a similar approach makes sense for animal experiments. In a study published this month in PLOS Biology, Daniel Strech, a bioethicist at Hannover Medical School in Germany, and colleagues investigated the idea of so-called animal study registries. They scoured the literature and interviewed nearly two dozen scientists to determine the pros and cons of such registries—and whether they would actually make a difference. Strech chatted with Science to discuss the group’s findings. This interview has been edited for clarity and length.
Q: What would these registries look like?
A: Before they start their animal experiments, scientists would need to document the details of the proposed study in an online database. This would include information like sample size, planned duration of study, how the experiments will be controlled, and the planned statistical analysis. A lot of these details are missing in scientific publications, which allows researchers to change their methodology after they begin a study to get the outcome they want.
Q: What are the advantages of such registries?
A: Animal researchers are supposed to avoid unnecessary and duplicative studies. If they are thinking about conducting a study, but a registry reveals that a similar study has already been done or is in progress, they might not do it. Or they may contact the team to try to set up a collaboration. The data could also help them refine their own study; if they’re trying to figure out what dosages of an antibiotic are most effective, for example, they can avoid dosages that another group has already tested.
Transparency is also a big advantage. Researchers might be less likely to register a poorly designed animal study if they know other scientists will see it, and that might stop them from conducting the study. And if you register a study with 20 mice but then only publish results with 13 mice, people are going to wonder what happened to the other seven mice. That could cut down on publication bias.
Q: What are the disadvantages?
A: One of the biggest differences between human registries and animal registries is the use of experimental therapies or compounds in animal studies that aren’t known to the wider scientific community. So a lot of people are concerned about the theft of ideas—that another, bigger lab could take your idea and do it quicker. Animal registries could get around this concern by instituting a confidentiality timeframe, where others couldn’t access your registry until after you publish your work. You could also limit the level of detail you give.
Some say animal registries would also add unnecessary regulations and paperwork. But to me it seems much less time intensive than all of the other things you have to do for this research—getting funding, approval, etc. Researchers are increasingly using digital versions of their protocol; you could just submit that.
Finally, some researchers worry that even though these registries are supposed to reduce the number of animals in biomedical research, their labs may end up using more animals because they feel they need larger sample sizes to conduct more robust research as part of these registries. Others say we’ll use fewer animals because overall the studies will be less wasteful. Either way, a lot of researchers are concerned that animal registries will tip off animal rights activists. I was surprised how often that concern came up. But nowadays you don’t need a registry to know which university is doing which research with animals.
Q: Will animal registries really have an impact?
A: Registries don’t directly decrease publication bias; they just shed more light on it. But they do provide a lot of incentive for researchers to eventually publish something. And registries should reduce redundancy regardless. In our study, people told us that, of course, they would check a registry to make sure a study hasn’t already been done.
With human registries, it’s been less than a decade since the FDA said all human clinical trials had to be registered. So it’s hard to tell how effective these registries have been. We also don’t know what the real publication bias was in human trials, so it’s hard to compare what has changed.
Q: Are you advocating for these registries?
A: When we started looking into this, we just wanted information. Now, I think the case is very clear that we need something like these registries. We need to design them as best as possible to address the interests of researchers, funders, and animals. But I’m still not clear what they should look like.
Q: What’s next?
A: If we want these registries to become a reality, we need a government agency like the FDA to mandate them. Journals could also require that a trial be registered as a condition for publication. Universities could advocate that their labs participate as well. In the meantime, groups like ours are working on a survey to determine how important the perceived strengths and weaknesses of these registries are to different stakeholders. It took decades from the first ideas of having a registry for human clinical trials to the reality in 2007, but I don’t think it will take that long for animal registries. I think we can skip some of those debates now, because we have solved some of the problems.