Should research ethics committees police reporting bias?
In: Kolstoe , S E , Shanahan , D R & Wisely , J 2017 , ' Should research ethics committees police reporting bias? ' British Medical Journal (BMJ) , vol 356 . DOI:10.1136/bmj.j1501
Progress through the application of science cannot be achieved if the results of scientific studies are not communicated appropriately. This is particularly relevant in medical research, where experiments are often conducted on human volunteers. Expensive new innovations cannot be produced without returns for investigators, but the distortion of the evidence base by long term withholding of data, either for financial gain or simply through negligence, can do real harm to patients. To tackle this problem the World Medical Association has set standards for the conduct and dissemination of research, but the sheer number of funders and researchers make it difficult even to identify trials that have been conducted, let alone to discover whether the results have been reported accurately. This problem, called reporting bias, has become increasingly prominent both in the public's consciousness and politically. Efforts to tackle the problem include trial registration, consensus based statements on how to report research, and audit attempts by funders, companies, and research organisations. However, these solutions lack coordination and examine only limited populations of researchers. What is needed is a larger and more systematic method for detecting this substantial ethical problem. The Declaration of Helsinki states that "any experiment involving human beings must be submitted to an independent committee for review, comment, and guidance," and these research ethics committees could hold the answer. For example, in England the Health Research Authority coordinates 66 committees that review the ethical aspects of all research on human participants conducted in the NHS (or subject to certain regulations). Researchers submit an application form and protocol for review, including a clear statement of the intended outcomes and how they will report the study. This process results in a substantial and comprehensive database of trials with human participants. Such databases represent an untapped resource, both for making the objectives and methods of planned trials in human participants available, and for determining whether the research has been reported.