Reporting of clinical trials in nursing journals: how are we doing?
Perhaps the most well publicised example of bad trials reporting is ‘study 329’. The trial involved the antidepressant paroxetine (Seroxat). The authors of the trial concluded that the drug was ‘generally well tolerated and effective’ in depressed adolescents. An independent reanalysis published in the BMJ in 2015 showed, in fact, that the drug was not effective (at least compared with placebo) and underreported an increase in harms (adverse events, specifically suicidal behaviour) in under 18 year olds who participated (Noury et al. 2015). Simply put, this study shows how bad reporting of clinical trials can put people's lives at risk.
Perhaps, as a consequence, public faith in science (not just the pharmaceutical industry) is declining (EuroScientist 2015). More optimistically, study 329 might be seen as a watershed moment in medical research. GSK (who manufacture Seroxat) have now signed up to the AllTrials campaign as have 88,381 individuals and 672 organizations (including Wiley that publishes JAN). Signing up to is easy, ensuring open and transparent reporting of clinical trials, it seems to us, is more challenging. For example, Reveiz et al. (2010) reported that only around a third of trials published in top ranked journals were registered. The authors compared a sample of 148 RCTs against CONSORT (http://www.consort-statement.org) standards and concluded that reporting was suboptimal (Reveiz et al. 2010). It is surely the key task of editors of scientific journals to ensure that trials are reported to the highest possible standard. We believe there is a need to examine the quality of trials reporting in nursing journals and have undertaken an audit of clinical trials published in JAN over the past five years (July 2011–June 2016).