Parallel Session G

Publication Bias

Moderator: Iain Chalmers, James Lind Initiative, Oxford, UK

Anne Brice and Iain Chalmers
James Lind Initiative, Summertown Pavilion, Middle Way, Oxford OX2 7LG, UK

Background:In an applied field of research such as health care reporting biases can result in avoidable morbidity and mortality and waste of resources. Reporting biases result principally from biased decisions to submit reports for publication by research sponsors and researchers. However, some journals also play a part by explicitly or implicitly discouraging submission of studies with statistically inconclusive results.  

Aim:  To assess the extent to which healthcare journals encourage or discourage authors to submit reports of scientifically robust research which has addressed important questions, regardless of the direction or strength of the results.

Methods:  We have selected for review journals covered by the MORE (McMaster Online Rating of Evidence) system. We have searched text on the websites of these journals to identify passages addressing the issue of reporting/publication bias.  

Findings: will be presented. 

 

Do medical journals help prevent publication bias by requiring trial registration?

Elizabeth Wager
Sideview, Princes Risborough, HP27 9DE, UK
in collaboration with the OPEN Consortium (Overcoming failure to Publish nEgative fiNdings)

Background:Registering clinical trials before they start can help reduce (or at least identify) non-publication and redundant publication. Trial registration can also prevent selective reporting (since trial designs on the register can be compared with published reports of findings). Some journals therefore refuse to publish trials that have not been properly registered. This initiative was led by the International Committee of Medical Journal Editors in 2004. However, not all journals require trial registration.

Aim: (1) To discover what proportion of journals require or encourage trial registration. (2) To discover editors’ reasons for not making registration a compulsory requirement for publication.

Methods:(1) Survey of journal instructions to authors. (2) Semi-structured interviews and qualitative analysis of findings.

Findings:will be presented.

 

Preventing publication bias: opinions of researchers and reviewers

Ana Marušić
Coeditor in Chief, Journal of Global Health; Department of Research in Biomedicine and Health, University of Split School of Medicine, Split, Croatia

Biases in study publication and outcome reporting are two major obstacles to evidence-based practice because they overestimate the effect of experimental treatments, can cause harm and are unethical. In recent years, the transparency of clinical trials has considerably increased with the establishment of public clinical trial registries and legislative changes in many countries. We explored the opinions of authors and reviewers of clinical research about their experiences with publication bias and potential solutions to the problem. The survey is planned within the framework of OPEN – Overcome failure to Publish nEgative fiNdings – grant from the FP7 HEALTH-2011 programme.