GBSI issues Reproducibility2020 challenge to stakeholders; Arturo Casadevall, Amy Herr, Joshua LaBaer, Brian Nosek, Timothy Simcoe and Richard Harris explore what ideas are working and where we go from here
WASHINGTON, D.C., February 17, 2016 – More than 70 leading scientists and industry leaders gathered last week in Washington, DC and via webcast to take on the problem of reproducibility in preclinical research that is a major concern in the biomedical community and continues to contribute to delays and increased costs of discoveries and development of treatments and cures. Global Biological Standard Institute (GBSI) hosted the discussion at its 2016 BioPolicy Summit—Research Reproducibility: Innovative Solutions to Drive Quality. The archived webcast is available now.
“Over the long-term, science is self-correcting, but short-term effects of irreproducible research include wasted time and resources, perpetuated errors in the literature and ultimately, mistrust of the biomedical research enterprise,” said Dr. Leonard P. Freedman, PhD, president of GBSI. “By raising the reproducibility rate through the expanded development and use of consensus-based standards and well-documented best practices, we would more effectively leverage the investment that funders—government, industry, and foundations—make in research, and yes—speed the development of new diagnostics, therapies and cures.”
Dr. Freedman introduced Reproducibility2020—a new GBSI challenge and action plan for the biomedical research community—to significantly improve the quality of preclinical biological research by year 2020 and encouraged participation among the attendees. “GBSI will devote its resources to affect change where its programs can make the greatest difference,” said Freedman. “We will work with partners to advance the broader agenda to improve the quality of scientific discovery in preclinical biological research settings, and keep you, and other stakeholders informed about movement and progress.”
Judith Kimble, PhD, University of Wisconsin, launched the discussion by outlining the problem and addressing some of the broader challenges around resolving reproducibility in research. NPR’s Richard Harris, steered the panel discussion with premiere life science thought leaders, Arturo Casadevall, Amy Herr, Joshua LaBaer, Brian Nosek and Timothy Simcoe, representing diverse perspectives from throughout the research community.
The panel reviewed progress of traditional avenues and new initiatives particularly in areas of transparency and training – sharing of methodologies; providing incentives and recognizing, such as with online badges, open access of data in journal publications; training in experimental design and reagent validation; and making biospecimen SOPs available on an online registry openly linked to related publications.
“We have a project together with GBSI to try to create what we’re calling a registry of protocols, so that when people deposit tissue samples, they would also publish how they collected the samples—what was their procedure,” said Joshua LaBaer, MD, PhD, Director, Virginia G. Piper Center at the BioDesign Institute at Arizona State University. “The idea is that, if from the beginning we can make their procedures transparent then we can at least start comparing and heading toward reproducibility.”
The group also explored more fundamental solutions and the potential benefits vs. detriments of the reproducibility pursuit, including de-emphasizing competition by rewarding curiosity, creativity and solutions; emphasizing performance measurement in place of competition; viewing error as opportunity when it instigates understanding; and revamping graduate education to change the foundation students are given and how adequately they are prepared to be scientists.
“Today in biomedical sciences…we often don’t train people on the basics of science…we train them primarily to be post-docs and not necessarily to be scientists,” said Arturo Casadevall, MD, PhD, professor, Department of Molecular Microbiology and Immunology at the Johns Hopkins School of Public Health. “One thing we are trying to do at Hopkins is to launch a new PhD program in the near future to change the way biomedical sciences students are trained. We call it ‘putting the Ph back in PhD’—integrating lab work within the bigger picture of philosophy of science. We will take students and train them in the basics – what is reproducibility? Then we’ll teach the anatomy of scientific error—an analysis of error and retractions. We don’t currently formally teach these things in the laboratory sciences and education is one area where we can make a big impact in the issues related to reproducibility.”
Harris opened the discussion to scientists and industry leaders in attendance adding new insights and issues to the debate, including getting more buy-in from universities and their administrations. U.S. medical schools have steadily shifted their financial responsibility to NIH (i.e., buildings, indirect costs, faculty salaries). Their competition for rankings, typically based on grant dollars, has become part of the problem that has contributed to an unhealthy environment.
“There’s great value in gathering great minds on a tenacious issue like irreproducibility,” said Freedman. “If we can foster a culture that drives more reproducible research, we will be able to accelerate the translation of research breakthroughs into life-saving therapies.”
Irreproducible preclinical research is a global, expensive, and increasingly well-recognized problem. Conservatively, 50 percent of published preclinical research is irreproducible, with an estimated annual cost of $28B in the United States alone.
About Global Biological Standards Institute (GBSI)
GBSI, a non-profit organization, is dedicated to enhancing the quality of biomedical research by advocating best practices and standards to accelerate the translation of research breakthroughs into life-saving therapies. For more information, visit GBSI.org; Twitter @GBSIorg.
Carol Miller at email: email@example.com or phone: 202-306-0130
Nancy Retherford at email: firstname.lastname@example.org or phone: 317-460-6838