Keywords: preclinical, animal-based, experimental design, messaging
Focus: Numerous exciting and apparently successful animal-based trials have not translated into successful clinical trials, either because promising claims for interventions could not be upheld, or because of failure to anticipate adverse effects in humans. Issues related to study design and bias have been implicated in poor reproducibility and translation of animal-based experiments. Common issues include failure to randomize or blind subject allocation and assessment, no formal experimental designs, P-hacking, and failure to report common confounders. Although reporting guidelines that could serve to alert investigators to these issues have been endorsed by leading journals, there is little enforcement and almost no change in standards over the past decade.
Discussion items: 1. How can statisticians increase awareness of the importance of good experimental design in preclinical research? 2. How can statisticians educate investigators in best scientific/statistical practice? 3. How can statisticians better disseminate their message?