Emphasizing the fact that clinical trials can influence treatment decisions, the recent AllTrials campaign has called for all trials past and present to be registered, and the results reported. As one of the initiating organisations behind this initiative, PLOS is committed to continuously developing and adopting policies that might help rectify publication bias in the trials literature. Following in the footsteps of PLOS Medicine, this campaign has led the editorial staff at PLOS ONE to re-evaluate how we handle unregistered trials.
Previously, in line with the International Committee of Medical Journal Editors (ICMJE) policy and like several other journals, we required all clinical trials that began after July 1, 2005 to be registered prospectively (before patient recruitment) in a publicly available registry approved by the World Health Organization (WHO). To stress the need for prospective registration, papers that did not fulfill these criteria were rejected, but this potentially prevented the results of some trials from being published. Thus while we still expect prospective registration to be standard practice, as of today, we will also consider retrospectively registered clinical trials to help ensure the medical community has access to the results of these trials.
These retrospectively registered clinical trials must however meet the following criteria:
1) The trial must be registered in a WHO-approved registry before submission to the journal
2) The authors must explain the reason for late registration within the methods section of the manuscript
3) They must include a statement confirming that all ongoing or future related trials are registered
As a reminder, when determining what constitutes a clinical trial, PLOS ONE follows the WHO’s definition: “a clinical trial is any research study that prospectively assigns human participants or groups of humans to one or more health-related interventions to evaluate the effects on health outcomes.” The ICMJE further clarifies that “health-related interventions include any intervention used to modify a biomedical or health-related outcome (for example, drugs, surgical procedures, devices, behavioral treatments, dietary interventions, and process-of-care changes). Health outcomes include any biomedical or health-related measures obtained in patients or participants, including pharmacokinetic measures and adverse events”.
The ICMJE also recommends that if scientists are unsure whether their trial meets the above definition, they should err on the side of registration.
We believe that while prospective trial registration remains the gold standard, this change in policy will be a positive step towards increased transparency and accountability and is in line with another recent initiative supported by PLOS ONE, the RIAT (Restoring Invisible and Abandoned Trials) initiative. This is also in accordance with the core values of the journal to publish all papers, including studies reporting negative results, provided they are technically sound and done to high scientific and ethical standards.
Image: by NHS Confederation on Flickr