Gøtzsche PC et al. (2007) Ghost authorship in industry-initiated randomised trials. PLoS Med 4: e19

In a Danish cohort study, Gøtzsche et al. have investigated the prevalence of 'ghost authorship' in industry-initiated clinical trials. They defined a 'ghost author' as an individual who contributed to statistical analysis or to the writing of the trial protocol or published manuscript without being identified as an author, as a member of a study group or writing committee, or in the acknowledgments section. By examining original full trial protocols and subsequent primary papers, the authors found evidence of ghost authorship in 33 of 44 (75%) industry-initiated trials that received approval in the period 1994–1995 (data published 1997–2002). This is in contrast to the results of self-reporting surveys on ghost authorship, which have estimated prevalences of 10–15%.

Statisticians were the ghost authors identified most frequently and most easily. Recent guidelines issued by the International Committee of Medical Journal Editors do not recommend that individuals who have performed statistical analyses be listed as authors unless they have contributed to other aspects of the paper, but do recommend that such persons be acknowledged. Gøtzsche et al. contend that the role of statisticians in analysis of trial data should be more clearly stated, both to improve accountability and to provide more information as to the role of industry in these studies.

The investigators conclude that ghost authorship is common in industry-initiated trials and that the practice serves a commercial purpose. They suggest that adherence to existing guidelines would reduce the prevalence of ghost authorship, and that more journals should state the individual contributions made by each named author of a manuscript.