Prevalence of Honorary and Ghost Authorship in Cochrane Reviews

Context To determine the prevalence of honorary and ghost authorship in Cochrane reviews, how authorship is assigned, and the ways in which authors and Cochrane editorial teams contribute.

Methods Using a Web-based, self-administered survey, corresponding authors for 577 reviews published in issues 1 and 2 from 1999 of The Cochrane Library were invited to report on the prevalence of honorary and ghost authors, contributions by authors listed in the byline and members of Cochrane editorial teams, and identification of methods of assigning authorship. Responses were received for 362 reviews (63% response rate), which contained 913 authors.

Results One hundred forty-one reviews (39%) had evidence of honorary authors, 32 (9%) had evidence of ghost authors (most commonly a member of the Cochrane editorial team), and 9 (2%) had evidence of both honorary and ghost authors. The editorial teams contributed in a wide variety of ways to 301 reviews (83%). Authorship was decided by the group of authors (31%) or lead author (25%) in most reviews. Authorship order was assigned according to contribution in most reviews (76%). The 3 functions contributed to most by those listed in the byline were assessing the quality of included studies (83%), interpreting data (82%), and abstracting data from included studies (77%).

Conclusions A substantial proportion of reviews had evidence of honorary and ghost authorship. The Cochrane editorial teams contributed to most Cochrane reviews.

Authorship in biomedical publications establishes accountability, responsibility, and credit.¹ A survey by Flanagin et al² found that 26% of review articles in 6 peer-reviewed medical journals contained evidence of honorary authorship (naming as an author an individual who does not meet authorship criteria) and 10% of review articles contained evidence of ghost authorship (failure to name as an author an individual who has made substantial contributions that merited authorship or an unnamed individual who participated in writing the article). Other studies^3-5 have confirmed that a high proportion of authors listed in the byline do not fulfill criteria for authorship.

The Cochrane Collaboration is an international organization that was established to help individuals make well-informed decisions about health care by preparing, maintaining, and ensuring the accessibility of systematic reviews of the benefits and risks of health care interventions.⁶ Cochrane Collaborative Review Groups (CRGs) support authors worldwide in preparing and maintaining systematic reviews according to a common methodologic framework.⁷ Other types of groups, devoted to methods, dimensions of health care, consumer perspectives, and geographic regions, also contribute to the Cochrane Collaboration's work. Cochrane Collaborative Review Groups cover most of the major areas of health care and are coordinated by an editorial team, which is composed of a CRG coordinator (who manages and supports a CRG on a day-to-day basis), a trials search coordinator (who is responsible for searching for studies that fall within the scope of the group and maintaining their details in a specialized register), and a group of editors led by a coordinating editor. The editorial teams edit and assemble completed reviews for inclusion in The Cochrane Library, a collection of databases published on CD-ROM and the Web and updated quarterly. Within The Cochrane Library, Cochrane reviews are published in the Cochrane Database of Systematic Reviews. A corresponding author is named for each review.

Because Cochrane reviews are published electronically, updated quarterly, and dependent on assistance from CRGs, they may generate practices of authorship and contributorship that differ from those generated by print articles. However, little is known about these practices.

The objectives of this study were to determine the prevalence of honorary and ghost authorship in Cochrane reviews, the ways in which authors and members of CRG editorial teams contribute to Cochrane reviews, and the methods by which authorship is assigned.

In March 2000, we invited the corresponding authors of the 577 Cochrane reviews published in issues 1 and 2 from 1999 of The Cochrane Library to complete a 29-item Web-based questionnaire. The questionnaire was piloted before the launch of the survey and was based on one previously used by Flanagin et al² in a study aimed at determining the prevalence of honorary and ghost authors in peer-reviewed biomedical journals. We used the software package SPSS version 9 (SPSS Inc, Chicago, Ill) to analyze the completed questionnaires.

Our criteria for authorship were based on recommendations of the International Committee of Medical Journal Editors (ICMJE)⁸ in March 2000, when the survey was launched. Three conditions defined authorship: (1) conceiving and designing the review or analyzing and interpreting the data (we considered the following contributions, common to the conduct of systematic reviews, to fulfill this criterion: conceived of or designed the review, designed the literature search, conducted the literature search, screened results of searches, assessed quality of included studies, abstracted data from included studies, performed statistical analysis, interpreted data, contacted authors of primary research or industry, or supervised the work of coauthors, if any); (2) drafting the review or revising it to make important changes in content; and (3) approving the final version before it was submitted to the editorial base of the CRG through which it was to be published in The Cochrane Library. Any author not meeting all 3 conditions was classified as an honorary author.

We defined a ghost author as a person who, in the opinion of the corresponding author, had made a contribution that merited authorship or who had assisted in drafting the review but was not listed as an author or mentioned in the acknowledgment section of the review. Corresponding authors were also asked to indicate, from a list of functions, the support received from members of the CRG editorial team (Table 1).

Corresponding authors for 362 of 577 reviews (63% response rate) completed the questionnaire. Of these 362 reviews, 13% had a single author, 49% had 2 authors, 21% had 3 authors, 11% had 4 authors, and 6% had 5 or more authors. In comparison, of the 215 reviews for which no response to the questionnaire was received, 23% had a single author, 32% had 2 authors, 16% had 3 authors, 10% had 4 authors, and 19% had 5 or more authors.

The 362 reviews contained a total of 913 authors. Of the 913 authors, 18 (2%) did not participate in conceiving and designing the review or analyzing and interpreting the data, 204 (22%) did not participate in drafting or revising the review, and 168 (18%) did not participate in approving the review before it was submitted. Of the 362 reviews, 141 (39%) had evidence of honorary authors. Fourteen reviews (4%) contained at least 1 author who had not participated in conception and design or conduct, 87 reviews (24%) contained at least 1 author who had not participated in drafting or revising the review, and 98 reviews (27%) contained at least 1 author who had not participated in approving the review before it was submitted.

Authors listed in the byline contributed to a variety of functions (Table 1); the functions contributed to most often were assessing the quality of included studies (83%), interpreting data (82%), and abstracting data from included studies (77%).

Thirty-two reviews (9%) had evidence of ghost authors, although from the information provided it was not possible to determine whether the contribution of the ghost authors merited authorship or whether a mention in the acknowledgment section would have sufficed. The CRG editorial team members were identified as ghost authors in 22 reviews (6%). Other ghost authors identified in 10 reviews (3%) included named individuals, lay reviewers, peer reviewers, or authors who had contributed to a previous version of the review but were not involved in the current version. Nine reviews (2%) contained both honorary and ghost authors.

Authorship was decided by the group of authors (31%) or lead author (25%) in most reviews. There was no formal mechanism invoked for deciding authorship in 28% of reviews. Authorship order was assigned according to contribution in most reviews (76%) and alphabetically by author surname in 9 reviews (2%). In 79 reviews (22%), a variety of other methods were used to assign authorship, including seniority of those involved, senior author last, responsibility for the work, internal discussion and agreement, and the stage at which the authors became involved in the review. The CRG editorial teams contributed to 301 reviews (83%) in a variety of ways (Table 1 lists specific functions).

Honorary authorship in Cochrane reviews was more prevalent than ghost authorship. The main reasons for authors being classified as honorary authors were that they had not participated in drafting or revising the review or had not approved it for submission to the editorial base of the CRG. The prevalence of honorary authorship in Cochrane reviews was greater than that reported for review articles in peer-reviewed medical journals,² although this may be at least partly explained by differences in the definition of honorary authorship used in the 2 studies. The previous study used a less strict definition of actual authorship, with an article defined as having an honorary author if the corresponding author did not fulfill the ICMJE authorship criteria, would not feel comfortable explaining the major conclusions of the article or review, or reported that a coauthor performed only one function and nothing else from a list of 17 activities, whereas in this study the authorship criteria were applied equally to all authors.

In most reviews, the group of authors or lead author decided authorship, and the order of authors was assigned according to the contribution they had made. The CRG editorial teams made substantial contributions to most Cochrane reviews, especially to conception and design, literature searches, and preparation for publication.

Our survey had limitations. The information was based on self-report by the corresponding authors on behalf of their coauthors. Also, the response rate to our survey was 63%, and we do not know if the reviews of nonrespondents differed systematically from those of respondents in terms of the prevalence of honorary or ghost authors or the ways in which authors and CRG editorial teams contributed to the reviews.

The rate of ghost authorship may be greater than we found, since our definition excluded the possibility that a person named in the acknowledgment section of the review could have made a contribution that merited authorship. Conversely, the rate could be an overestimate, since we included among ghost authors those individuals who had contributed to a previous version of the review but were not involved in the current version. Such persons should not qualify for authorship if they do not contribute to the update of the review, in the same way as authors who have not been involved in a follow-up of a previously published trial do not qualify for authorship in the update, although much of the material and text may be the same.

Steps have been taken by the Cochrane Collaboration to improve the reporting of contributions to reviews. Since our survey was undertaken, a section in Cochrane reviews has been created where the names and contributions of all individuals who have contributed to a Cochrane review (including the editorial team of the CRG) can be described. The Cochrane Reviewers Handbook⁷ gives guidance on this, providing a list of potential contributions adapted from Yank and Rennie.⁹ The Handbook⁷ now also recommends that authorship criteria for Cochrane reviews should be based on the recommendations of the ICMJE. Adherence to these recommendations should help reduce the prevalence of honorary and ghost authorship and encourage transparency in the reporting of contributions. The Cochrane Collaboration might also wish to consider implementing a version of the disclosure policies already adopted by some journals,¹⁰ which require authors to describe their contributions in their own words or to complete a checklist on authorship criteria and contributions.