Intersection of Systematic Review Methodology with the NIH Reproducibility Initiative

In a landmark 2005 paper published in PLoS Medicine, Ioannidis posited that " most current published research findings are false " (Ioannidis 2005). Consistent with this opinion are reports that drug develop ment has been hin­ dered and many clinical trials wasted based on published findings from pre­ clinical studies that with further effort could not be reproduced (Begley and Ellis 2012). The National Institutes of Health (NIH) recently outlined a sweeping set of initia tives to address the lack of reproducibility of research findings (Collins and Tabak 2014). In this editorial we touch on current efforts to address the research repro­ ducibility problem and propose that systematic review methodologies, which are being developed to assess confi­ dence in the quality of evidence used in reaching public health decisions, could also be used to improve the reproducibly of research. Reports and editorials in the biomedical literature have increas­ ingly drawn attention to a disturbing lack of reproducibility of pub­ lished scien tific findings. Although poor reporting of key aspects of study methodology clearly contributes to the problem, other factors such as study conduct, may be equally or more important (Begley and This situa tion has prompted actions by both the private and public sectors. These include the private Reproducibility Initiative collabora­ tion between Plos One 2013), which among other projects is attempting to replicate key find­ ings from the 50 most impactful studies published in the field of cancer biology between 2010 and 2012. A major public effort is the NIH Initiative to Enhance Reproducibility and Transparency of Research Findings, which seeks to increase community awareness of the repro­ ducibility problem, enhance formal training of investigators in elements of proper study design, improve the review of grant applications, and increase funding stability for investigators to enable them to use more appropriate and complex study designs (Tabak 2013). One planned activity of the NIH initiative is to develop a pilot training module on research integrity as it relates to experimental biases and study design. The intention is to provide specific guidance for researchers to improve the quality of their research publications by increasing their awareness of research practices that may affect the validity of their study findings. This guidance could also be used to improve both the grant proposal and journal peer­review stages to ensure more systematic and rigorous evaluation of both proposed and completed studies. Hoojimans and Ritskes­Hoitinga (2013) recently …