When I sat down to write this blog post, I opened a blank page and just started typing – my tried and tested method for writing any piece of work. I sent the document for review and the feedback I received was something along the lines of “Yes, this is a blog article – but the execution could have been better”. Fortunately, for this piece, I can go back and start again from the beginning. I was directed to our Comms Template to plan my article more rigorously, before then using that framework to write a new blog post. You won’t see that first piece of work; it is destined for the recycling bin. The second (or third) attempt will be the one on your screens now.

Up to 40% of research involving animals does not make it to the public domain

Taking this story for the parable that it is, the aim of the work we do as scientists is to advance knowledge in our research field and share our findings. But a major issue in preclinical animal research is that a lot of our findings never get published – that’s data going straight in the recycling bin. How much is difficult to quantify, but research from the Netherlands and Germany suggests that up to 40% of research involving animals does not make it to the public domain.

In the literature that we do see published, a vast majority of it does not include vital information to interrogate the rigour and reliability of the experiments presented. The reason is likely a combination of authors not being aware of the need to report on such criteria alongside an unwillingness to present some of the flaws in the design of the experiment lest it hinder the already precarious chances of publication. As reviewer two might say, “the execution could have been better”.

Thankfully, guidelines for reporting experiments have been created to help researchers. The ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines were developed in 2010 by the National Centre for the Replacement, Refinement and Reduction of Animals in Research (NC3Rs) to help scientists maximise the quality of published research and to promote the full and transparent reporting of experiments involving animals.

However, even as awareness of ARRIVE has grown over the years, the 2019 IICARus study demonstrated that requesting the essential information at the publication stage does not altogether improve transparent reporting. In 2020, the guidelines were updated and ARRIVE 2.0 now includes an Explanation and Elaboration (E&E) for each of the checklist criteria. These explain the background and rationale of the guidelines and provide examples of good reporting from the published literature. But, experimental rigour needs to be considered throughout the research cycle (from the funding application to publication) and the framework for reporting a study should be fully considered at the study design phase.

The world’s largest public funder of biomedical research, The National Institutes of Health (NIH), served notice earlier this year of the expectation for researchers to use the ARRIVE guidelines when conducting and reporting animal experiments. This endorsement is a testament to the significance of the ARRIVE guidelines in maintaining scientific integrity. When designing your next in vivo experiment, use the ARRIVE website to consider how to include all the information necessary to showcase your results. Unlike this blog article, where I could start afresh, restarting an experiment from the beginning due to sub-optimal design is wasteful (in terms of research funding, scientists’ time and animals’ lives).

Masking is a collaborative effort

An example of information that is often missing from the literature and one that cannot be conducted after a study has finished is item five of the checklist: masking, also known as blinding. This is where investigators performing the experiment or those measuring or analysing data don’t know which treatment the animals are receiving or have received at different stages of the study. Masking is an essential mitigation to unconscious bias (where we expect to see a particular outcome in our study) and we should mask wherever it is possible to do so. It may not be possible to mask the investigator at all stages of an experiment. For example, if a scientist wants to compare drug treatment in two mouse genotypes, they may be unable to mask the researcher performing the experiment to the genotype of the animal if their phenotype (visual identity) is different – but they could improve the experimental design by masking the researcher during analysis of the final dataset by independently applying a code to the results. Masking is a collaborative effort and the NC3Rs have produced further guidance on implementing masking for your next study.

One of the most notable additions to ARRIVE 2.0 is the addition of protocol registration. Researchers are now encouraged to preregister their experiments, which involves documenting their research protocols and methods in advance. This helps to mitigate against publication bias, where published literature includes disproportionally more ‘positive’ or statistically significant results than should be expected and encourages the publication of all research data. Publication bias is not unique to any scientific discipline, but the biology and biochemistry fields are some of the worst offenders.

This bias for data deemed to be more important or impactful presents a number of problems. Not least is that unreported ‘null’ research is needlessly repeated by multiple scientific groups, wasting time, money and animals’ lives. On top of this, the true effect size of a new intervention may also be over-estimated. For example, compounds to treat stroke or neurological disease are presented in the literature as much more efficacious than the clinical reality, limiting the true translatability of preclinical studies. Preregistering a study protects scientists from selectively reporting results (unconscious bias makes this is all too easy to do for even the most well-meaning researchers) and can improve the transparency of animal research. It can be done directly with a journal, using the Registered Report article type, which has the added benefits of a guaranteed publication regardless of the results. Peer review of the methods and hypothesis are also conducted prior to the study start, allowing researchers to incorporate expert feedback and improve their experimental design in the process.

Having ARRIVE to hand when planning your next experiment is key

Using ARRIVE 2.0 at the earliest stages of the research pipeline helps scientists to design, perform and report the highest quality research. Over a thousand funders, journals and publishers require that studies are planned and reported in accordance with ARRIVE, to improve confidence in scientific research and results. Having ARRIVE to hand when planning your next experiment is key, and resources like the pocket-sized Z-cards are available to order directly from the NC3Rs for an easy desk-side reminder of essential design and reporting standards. Going forward, the NC3Rs are working on creating an AI tool to automatically check scientific text for adherence to the ARRIVE Essential 10, making it easier to identify missing information as early as possible in the publication process.

Top image by ARRIVE