Copyright 2004 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2004
The article by Hartling et al1 in this issue of the ARCHIVES joins a growing number of studies identifying the underlying reasons for publication bias. As with other studies, what is demonstrated here is that the root cause of nonpublication lies with the investigator, not with the editor or the work itself.2- 6 More than 1 in 4 of the randomized controlled trials reviewed were not published, and, of those, a surprisingly low 17% were actually ever submitted for publication. On the surface, this newest study seems to be a commentary on the busy lives of physician researchers: research does not get published simply because there is not enough time. Additionally (and sadly), work does not get published because of difficulties with co-investigators. As discussed by the authors, one of the problems with nonpublication is the potential impact on clinical decision making. Clinical decisions based only on the published literature may have questionable validity. Evidence of the consequences of selective reporting and underreporting on clinical decision making has been documented.7 While this is clearly a crucial problem, there is an additional implication of nonpublication directly related to the ethics of human subject research. We must consider nonpublication essentially as a breach of contract between the researcher and the participant.
Dowd MD. Breaching the ContractThe Ethics of Nonpublication of Research Studies. Arch Pediatr Adolesc Med. 2004;158(10):1014-1015. doi:10.1001/archpedi.158.10.1014