It is worth remembering that the primary purpose of clinical research is to improve the health of individuals or a population. There is general agreement that the gold standard for clinical research, particularly for new treatments, is the well designed, adequately powered randomized controlled trial (RCT). An area that has received less attention is how those trials are reported and disseminated to healthcare professionals and policymakers who directly or indirectly influence patient care as well as the potential consequences of suboptimal trial reporting. In this issue of the Journal, Rios et al. (1) describe the quality of RCT reporting in three general endocrinology journals (The Journal of Clinical Endocrinology and Metabolism, Clinical Endocrinology, and the European Journal of Endocrinology). In this study, all treatment or prevention RCTs published in 2005 and 2006 in those journals were identified and independently reviewed by two investigators using established criteria. The criteria, developed and published by the Consolidated Standards for Reporting Trials (CONSORT) group (2), assess the fundamental design and analytic features that constitute high quality reporting of RCTs. Two primary outcomes were considered, a 15-point overall reporting quality score (OQS) based upon the presence or absence of 15 elements (such as a clear description of the inclusion criteria, precise details of the interventions, and inclusion of a flowchart of participants at each stage of the study) and a key score based upon three critical elements: allocation concealment (i.e. ensuring that investigators and subjects remain unaware of upcoming assignments), blinding, and use of intention-totreat analysis. The results of the study were not encouraging; the mean overall reporting quality score of the reviewed trials was only10 of 15 points (interquartile range 2), and the rates of adequate reporting of allocation concealment (10%), blinding (20%), and intention-to-treat analysis (16%) were shockingly low. An important detail is that of the 89 trials reviewed, 71 were published in the journal with the highest impact factor and widest readership (The Journal of Clinical Endocrinology and Metabolism). The authors correctly conclude that the reporting of RCTs in those general endocrine journals is suboptimal. Although the study appears to be rigorously designed and conducted, Rios et al. (1) are also guilty of reporting deficiencies, which may or may not reflect design flaws. For example, it is not stated whether the investigators reviewing and assessing trial quality criteria were masked to the identity of the authors or journal that published the trial. Such masking reduces the likelihood of bias. In fact, criteria similar to CONSORT but designed to judge the quality of reporting for observational studies have been published (3). Furthermore, as noted by Rios and others (4), inadequate reporting does not necessarily imply inadequate trial design and conduct, although it seems likely that that the two are correlated. Although discouraging, the results of the Rios study are not surprising. Using a variety of criteria, others have found that RCT reporting is suboptimal in journals devoted to surgery, nursing, and general medicine (5–7). It is likely that endocrinologists are no better or no worse than other specialists when reporting RCT results, and it should be acknowledged that the study by Rios et al. (1) did not include important endocrine trials reported in the New England Journal of Medicine and Journal of the American Medical Association, which have adopted the CONSORT guidelines as a requirement for publication. Who is to blame for the poor quality of RCT reporting in general endocrine journals? Although authors should strive to submit the best possible manuscript possible, journal editors are ultimately responsible for the quality of the manuscripts they publish. In turn, editorial decisions are strongly influenced by external reviewers. Both editors and peer-reviewers must better equip themselves with the necessary skills to identify reporting deficiencies and demand they be addressed before publication. Fortunately, there is a proven method to improve the qual-
Read full abstract