Level of bias (References) | Examples | Current ICMJE protection (section of December 2015 Recommendations) |
---|---|---|
Academic | • Clinical trials: what studies to undertake, whom to recruit as partners, whether and how to publish. • Publications: number, theme, placement, scheduling and authorship of articles. | • Limited options available to ICMJE. • Clinical trial registration (III L). • Duplicate/prior publications (III D). |
• Clinical trials: design, conduct, discontinuation, analysis and interpretation. • Data misrepresentation, fabrication and falsification. • Review criteria and literature search. • Insertion of commercial “key messages.” • Language and rhetoric. | • Limited options– heavily reliant on peer review. • Contributor listings (IIA 1). • Stringent author accountability (IIA2). • Access to data (IIB, IIB 2). • Authors’ right to publish. • Reporting guidelines (IV A2). • Commercial supplements and series – journal’s editorial control maintained (IIIG). • Rules on correction or retraction of inaccurate or fraudulent data (IIIB). | |
• Spinning the sum effect of authorship, contributorship, acknowledgements, text, labeling, etc. • Depends on what readers perceive, not merely what is disclosed. • Advocacy-based marketing foregrounds recruited academic authors while obscuring the proprietary role of commerce using small print, vague language and omission. | • Extensive options – but poorly developed. • Ban on ghost authorship (II A2). • Funding source, trial identifier listed visibly in Abstract (IV A3). • Contributorship, interest declarations and “Sources of support” sections provide limited benefits but involve small print (IIA, IIB, IV A3). • Declaration of product being marketed – but only for supplements (IIIG). |