Earlier this year, we (Nikki Osborne, Director of Responsible Research in Practice and Jane Alfred, Director of Catalyst Editorial) conducted a survey of the research community to explore researchers’ awareness and expectations of, and training provision in, the reporting of animal (preclinical) research. As many of you reading this blog will be aware, the inadequate and incomplete reporting of preclinical animal research has been linked to a failure to reproduce many reported animal experiments (see for example, Kilkenny et al, 2010 and Avey et al, 2016). This lack of reproducibility has far reaching implications for the preclinical research community and for improving our capability to understand, diagnose and treat medical conditions and disease.
In an effort to address this problem, the research community has devised various standards and guidelines to improve the planning (the PREPARE guidelines) and the reporting (such as the ARRIVE guidelines and the HARRP guidelines) of animal research. But while journals encourage authors to use these guidelines few, if any, as yet mandate them. And indeed recent studies show that animal research reporting guidelines alone are not enough to fix the problems associated with irreproducibility; we need to pay more attention to how experiments are designed and planned as well.
Our intention in sending out this survey, therefore, was to identify whether there are any unmet training needs in this area and if so what they are. And from this survey, we uncovered some unexpected observations that we feel are interesting in their own right and thus worth sharing more broadly through this blog.
But first, some background information about the survey. We sent a link to a mix of our existing clients and contacts, who work in a range of relevant research organisations and scientific societies. We encouraged the survey link to be shared by our contacts to help disseminate it to relevant individuals and organisations across Europe. As a result, we received a total of 113 responses in just over two weeks by the time the survey closed.
Most survey responses were from individuals (87%), with the remainder responding on the behalf of an academic institution (5%), research establishment (5%), or commercial organisation (3%). All respondents confirmed that their research involves animals, animal- derived materials and/or animal-derived data. Looking at respondents’ roles, the largest individual categories were group leaders (23%) and early career researchers (22%). Training officers and co-ordinators were the next highest responding group (13%), followed by policy or research directors (6%) and veterinarians (1%). The remaining 30% of respondents overall identified themselves as fulfilling two or more of these roles.
It is perhaps expected that nearly 90% of the survey’s responders told us that they are aware of the issues concerning the reproducibility of biomedical research, given the much needed attention now being given to this problem. However, only 83% of responders are aware of issues concerning the quality of animal research reporting and only 83% are aware of issues concerning the quality of biomedical research reporting more broadly. This tells us that 17% of the survey’s responders are unaware of these research reporting issues. More concerning still is that only 80% of the responders are aware that issues to do with animal research reporting are linked to the ability to use preclinical data in translational research, meaning that 20% are not.
When we asked our survey responders to share with us ‘other concerns’ they might have about animal research reporting, several key issues were raised, including publication bias and the selective reporting of animal research, sample sizes, the lack of proper controls, and other issues relating to study design and statistical analysis. One responder specifically commented on a lack of training on animal research in academic programmes. These areas are highlighted as training requirements within the new ‘curriculum for the use of animals in research’ published by the British Pharmacological Society for undergraduates.
It also surprised us to learn that only 38% of the survey’s responders receive training on how to report animal research data and on how to write research articles about animal data. 11% of responders told us that they receive no training on the areas we queried them on in the questionnaire. This tells us that there is a significant gap in training provision in animal research reporting. And it is a gap that the survey responders are keen to fill, with 77% of responders stating that they would like their organisation to provide them with training on animal research reporting and with 74% telling us that such training was relevant to their role.
Journals, research funders and scientific societies are proactively working together to tackle the culture change that is required to improve both the reproducibility and reliability of biomedical research, and the reporting of animal research. We also have a moral obligation to maximise the benefits of animal research and to reduce the number of poorly performed, poorly reported, or unnecessary studies. So surely now is the right time for research establishments to acknowledge their training responsibilities to support this culture change by providing training on this vitally important research output.
The expectation is clear, and the appetite for training to improve research reporting within the laboratory animal science community is there – so who would like to work with us to make this happen?
Authors: Nikki Osborne , Jane Alfred