An eye-popping $ 28 billion is spent in the US each year preclinical research that can not be replicated by other researchers. This is the conclusion of a provocative analysis published today in part by economists that are based on previous studies of error rates in biomedical studies.
Meanwhile, the National Institutes of Health (NIH) today issued new criteria for the review of grants aimed at strengthening the reproducibility of research funded by the NIH.
lead author of the new price tag for the reproducibility said it is intended to stimulate discussion. "We are showing the economic cost, but we also try to promote solutions. It's really the message of the paper, "says the biologist Leonard Freedman, president of the nonprofit Organic World Standards Institute (GBSI) in Washington, DC, from the viewpoint in PLOS Biology .
One expert, however, is skeptical about the figure of $ 28 billion, saying it may exaggerate the extent of any problem.
To come up with the number, economists Freedman and Iain Cockburn and Timothy Simcoe Boston University combed the literature for two dozen studies that have attempted to quantify the number of biomedical materials are defective due to specific problems such as contaminated cell line. Looking across these data, they estimate that 53% of all pre-clinical studies have errors which means they are not reproducible. The most common reasons included problems with reagents and reference materials (36%), the design of the study (28%), data analysis and reporting (25%), and laboratory protocols (11%).
The 53% is roughly comparable to a handful of studies "top-down" that tried to reproduce a set of conclusions, says Cockburn. For example, a widely cited analysis by Amgen revealed that only 11% of the 53 documents preclinical cancer could be replicated in the laboratory of the company. Other studies have shown a rate closer to 50%.
Hence, the calculation of the economic impact was simple. Crossing the 53% to a "conservative" 50% reproducible estimate, the researchers multiplied by $ 56 billion a year that the NIH and other backers of US public and private funds dedicated to preclinical research. This provides $ 28 billion in irreproducible preclinical research.
Does this mean the money is wasted? Not exactly, say the authors. Instead, they say that the reproducibility would fight the funding goes much further. "The time has come to invest more, not less, with a relatively small share of this investment to improve the reproducibility rates," said Freedman.
Other solutions, researchers should receive better training in the design of the study, and sellers and scientists should sell and use only validated reagents, PLOS Biology commentary concludes. GBSI will encourage the research community to establish accepted standards for authentication of cell lines.
NIH also taking steps to improve reproducibility. in an announcement today, the agency describes four new criteria that grant their authors will be invited to consider starting in January . They include next the strength of the scientific premise that the proposed study is based on; the rigor of the study design; the review of the sex proposal of research animals or human subjects; and if the reagents have been authenticated.
Although he has not seen the details of the microbiologist PLOS Biology comment, a scientist who has studied why the papers are retracted said it "sounds sensational." Ferric Fang of the University of Washington, Seattle, wonders if it is possible to extrapolate from some previous studies on the reproducibility and concluded that 50% of pre-clinical research can not be reproduced. This is partly because the inability to reproduce a specific experience is not the same as not to reproduce that found in different conditions; this result is most common but does not mean the original finding was erroneous. "To say that 50% of research funds are wasted is ridiculous and unnecessary," says Fang.
0 Komentar