The question of how quality of preclinical data is connected to patients in clinical trials is addressed in this article by Yarborough et al. This discussion is certainly overdue as the ethical implementations of doing tests in human subjects without adequate preclinical justification are not bearable for an advanced civilisation of the 21st century.
Yarborough and colleagues describe several issues in the context of preclinical research and look for sources of these problems but also – and this is new – for how preclinical research outcomes are communicated to ethical review boards and patients. The main issues identified for preclinical research are publication bias towards positive data and experiments that have not been sufficiently powered, blinded or replicated. These topics and potential consequences are not adequately communicated to research ethics committees and certainly not to the patients: research lacking rigor and reporting results selectively is not fit to either efficiently develop novel therapeutic strategies or assist research ethics committees (RECs) to weigh harms and benefits for patients in a meaningful way.
Ultimately, the responsibility to provide high quality research data lies with the bench scientists themselves. Thus, the authors conclude that – given the existence of many guidelines and recommendations – it should now be in the interest of all stakeholders to assure that these recommendations are followed.

LINK