The success of clinical studies depends on the reproducibility of preclinical research results (see J. Kimmelman and C. Federico Nature 542, 25–27; 2017). We propose a three-tier principle of transparency, replication and triangulation that should be achieved before publication, to ensure that the results warrant further study in preclinical and clinical trials.
Transparency focuses on the availability of complete and clear information about experimental methodology. This must be sufficient to allow the published study to be replicated under the same conditions by other investigators, with essentially the same primary outcomes.
Replicability should be tested by the original researchers and/or by others in the same laboratory, and confirmed using different samples or specimens. Ideally, an unrelated lab should perform independent replication based on the reported methods.
Triangulation confirms the study's central findings or models using different methodologies and experiments, ensuring that measurements converge from different experimental perspectives. For example, Robert Edwards and colleagues studied maturation of in vitro fertilized embryos in rats, mice and hamsters before predicting the technique's success in humans.
In our view, fulfilling all three tiers is manageable and essential. In the long term, the guaranteed quality of the results will more than justify the delay to publication.
Related links in Nature Research
About this article
Cite this article
Balas, E., Ellis, L. Three-point plan for reproducibility. Nature 543, 40 (2017). https://doi.org/10.1038/543040d
This article is cited by
Pre-clinical animal models are poor predictors of human toxicities in phase 1 oncology clinical trials
British Journal of Cancer (2020)