Raising the research quality bar
Reproducibility isn’t something that can be solved without considering the bigger research picture. So as part of efforts to improve the quality of research, we’ve collected tips and resources – relevant to each stage of the research process – from across the MRC community to help. Isabel Baker reports.
Methods are us
Good science needs good methods. Good methods ensure that health research and policy are built on the best possible evidence. Using robust, bullet-proof methods that are reliable and repeatable can also improve efficiency. Efficiency is important, as it’s not just taxpayers’ money at stake; valuable samples from humans and animals can often be used only once, and time donated by volunteers is precious.
Making the biomedical and healthcare research process more efficient is the job of the MRC-NIHR Methodology Research Programme. It encourages uptake of best practice by funding, and writing, accessible guidelines.
Recent examples are a framework to help researchers with stratified medicine research and guidelines for choosing effect size for clinical trials. The programme also supports studies to improve new methods across all research areas, to ensure new evidence is accessible to as many people as possible.
As well as improving methods, it’s important to learn how to apply existing ones well. A new annual five-day residential workshop on ‘Advanced Methods for Reproducible Science’, co-led by the Universities of Bristol and Oxford, aims to teach early-career researchers skills and methods to improve the quality and reproducibility of their work. It will run again in 2018 and 2019.
PhD student Monika Frysz, of the MRC Integrative Epidemiology Unit, attended the 2017 workshop. She learnt not only about the scale of the problem of reproducibility but also about initiatives and tools in place to help tackle it.
“One of the most beneficial sessions included practical tools for code sharing, software version control and introduction to a free software for classical and Bayesian statistics. The course was filled with fascinating talks and plenty of time for discussion with like-minded scientists!”
The devil’s in the design
“But we’ve always done it like this!” just won’t cut it. If you want funding, we need to understand what you’re proposing to do and, most importantly, why you’ve designed it that way. Last year the MRC added extra space to our grant application form for this reason. We’ve now updated our guidance to make it clear exactly what we need from you.
It’s your chance to convince us, and your peers reviewing grant applications, that your research design is robust. The simple message is: if you can’t tell us why you’re doing what you’re doing, we can’t fund your research. Check out the updated guidance for applicants (Section 188.8.131.52). There’s extra guidance for applications involving animals in there too (summary at Section 4.1.8).
There are free tools out there to help. The NC3Rs Experimental Design Assistant is a website to support researchers planning animal experiments. And the Systematic Review Facility, developed by the CAMARADES systematic review group and funded by the NC3Rs, is an online platform to help with systematic reviews of pre-clinical studies.
Do you have confidence in the quality and origin of your research materials? This is important to help nip reproducibility issues in the bud. We’ve supported lots of useful resources and facilities covering animal, cell and human research.
Study mouse genetics? MRC Harwell has more than 3,000 different mouse strains available for researchers to use and add to. Mice and data stored as part of the International Mouse Phenotyping Consortium are freely available for researchers. This project is aiming to ‘knock out’ and define every protein-coding gene in the mouse genome. And their bespoke genetically-modified mouse programme uses the latest advances in gene editing techniques to create mice strains for research.
Use stem cells in your research? Earlier this year the UK Stem Cell Bank at the National Institute for Biological Standards and Control released European-approved, quality-controlled stem cell lines. These aim to support development of new cell-based medicines to help bring therapies to clinical trial.
Work with human participants, their tissues or data? In an ever-evolving regulatory environment it’s important to keep up to date. The MRC Regulatory Support Centre provides support with navigating the complex regulatory, ethics and governance landscape.
Sharing your data and results speeds up the pace of science, helping to advance biomedical research and healthcare. This means registering your clinical trials, publishing your results in open access journals and allowing other researchers access to your data.
All types of clinical and public health intervention studies involving human participants can, and should, be publicly registered, not just those defined as clinical trials.
All types of studies can be registered on the Open Science Framework. In the last five years, 94% of MRC-funded clinical trials were registered but we need your help to increase this number.
Preprints are another good way of sharing science and knowledge across the world – we accept them in grant applications and five-yearly funding reviews. Find out how they can help your research by reading how others are using them.
We’ve had policies on open access and data sharing for a while now and expect all researchers to follow them. A good example is the MRC Cognition and Brain Sciences Unit’s open science web page. So if you haven’t already, open up – and tell us how you’re doing it so we can pass the knowledge on!
Know of a handy tool or tip? Share it with us @The_MRC
Visit our webpages to find out more.