[The danger of unpublished trial results]

Abstract

Unpublished and selectively published trial results may lead to distortion of the effectiveness and the profile of side effects of interventions. This jeopardizes patient safety and leads to unnecessary costs. Between 25 and 50% of all clinical trials remain unpublished. Primary outcomes, as originally defined in study protocols, often differ from those in the final publication. Positive trial results are almost 3 times as likely to be published as negative results. Commercial sponsorship is strongly associated with the publication of results that promote the interests of that sponsor, but non-publication and selective reporting is also frequent among non-commercial trials. Existing solutions, such as the requirement for prospective registration of trial protocols and the mandatory publication of trial results within 1 year after completion of the trial will only be successful if their compliance is more strictly monitored.