In this case, the named individuals meet full criteria for authorship, complete THE JOURNAL's authorship forms (which includes indicating responsibility and specific contributions, disclosing conflicts of interest, and transferring or waiving copyright), and have their specific contributions as authors published at the end of the article.⁴ The other group participants, who do not meet full authorship criteria, also may be listed as members of the group with their contributions or roles also designated (eg, investigators, study coordinators, members of the steering committee, members of the data and safety monitoring board).

Another approach involves identification of a writing group or committee in the byline as the group of authors writing on behalf of the research group:

Writing Group for the Women's Health Initiative Investigators

The individual authors constituting the writing group meet full criteria and requirements for authorship as stated above, and their names and contributions typically appear in the affiliation footnote on the first page of the article or are listed at the end of the article.¹ The names and roles of the other group participants also generally are listed at the end of the article.

However, some authors and research groups might prefer that only the group name appear in the bylines of their articles to emphasize the collaborative nature of the effort. In this case, another option is for the byline to list only the group name followed by an asterisk:

The DAVID Trial Investigators*

The asterisk corresponds to listing of the research group, which must include a specifically identified list of authors or writing group members whose names and contributions appear in the affiliation footnote on the first page or at the end of the article.⁵ This allows for group name recognition and formal association with the specific article, while maintaining the ability to identify named accountable authors. As with other options, all persons identified as authors or as part of a writing group must meet full authorship criteria, complete authorship forms, take responsibility for the article, and have their specific contributions as authors published at the end of the article.

In another possible, but rarely used format for group authorship, all members of the entire research group may be designated as authors, with only the name of the research group appearing in the byline. However, such an approach requires clear justification that each group member meets the full criteria and requirements to truly merit authorship.

All authors are required to indicate their contributions to the work,⁶ report their conflicts of interest,⁷ and transfer or waive their ownership of copyright for the work.⁸ Individuals who do not fulfill authorship criteria should not be identified as authors^9-10 simply by virtue of being a member of the research group, or performing some activities such as acquisition of funding, enrolling patients, assisting in manuscript preparation, or supervising study personnel.³ The names of members of the research group who have made these or other contributions to the work, but who do not qualify for authorship, should be listed at the end of the article, with a description of their role.

The publication of group names in bylines has been reported to be problematic for bibliographic databases, such as those developed by the National Library of Medicine (NLM) and the Institute for Scientific Information (ISI).¹¹ Both institutions are working to improve their ability to capture and cite the names of research groups that appear as authors and to identify individually named authors who are part of groups. Until recently, however, articles that listed only a group name in the byline may not have been identified in standard bibliometric searches and may not have been cited by other authors.^11-12 For example, group authors may have been added to the end of the title in the NLM index and the author field may have appeared as anonymous or blank (ie, "no authors listed"). In such cases, a PubMed search for articles using the research group or trial name as a search term might produce citations to a specific article. However, a search using the name of the first author in this group might not produce a citation to the same article. Moreover, the actual PubMed citation might indicate "no authors listed," and if the trial name was not part of the article title, the citation may not include any mention of the group at all. In the ISI database, the author field may have been listed as "anon" when no individual author was named in the byline. Thus, others may not be able to search for and capture citations that do not include the names of individual or group authors and may fail to cite them in subsequent papers.

Even when actually cited though, articles referenced by different means (ie, listing the group name, the first individual author, or a writing group) can create problems for bibliometric databases and may result in inconsistencies and inaccurate citation "weight" given to articles by ISI.¹³ For example, authors have cited the report of initial sequencing of the human genome published in Nature¹⁴ in 2001 by listing the International Human Genome Sequencing Consortium or by listing the first author (ie, ES Lander, et al). The ISI erroneously identified these types of citations as 2 different articles, and thus reported an undercount of total citations to the article.¹³ This became apparent when editors at Nature noted that the human genome sequence report published at the same time in Science¹⁵ was ranked by ISI as a highly cited paper and the comparable Nature article was not included on the high citation list.¹³ The Science article byline lists the individual authors by name (ie, J. Craig Venter, Mark A. Adams, Eugene W. Myers, et al).¹⁵ However, the Nature article lists the group name in the byline (ie, International Human Genome Sequencing Consortium*) and provides a partial list of named authors on the second page of the article that corresponds to the asterisk in the byline.¹⁴

To address these concerns, the ISI is working to link citations to articles whether the group name or first author in the group is cited. The NLM has designated 2 specific author fields: one for individually named authors and a second for group names. In addition, the current version of the NLM Indexing Manual indicates that any individuals who are part of a group and are listed explicitly as "authors" or members of a "writing committee," at any point in the article, also will be listed as well in the NLM author field.¹⁶ When such indexing is fully functional, PubMed citations to specific trials published after 2001 should be identified when searching by either the group name or an individual author's name, and citations to these articles should no longer appear as "no authors listed" provided journals indicate the name of the group, the names of the individual authors, or both.¹⁶

The names of research groups and individual authors should be formally associated with published articles as well as the bibliometric citations to these articles in online databases—and thus easily recognized and retrieved. Accordingly, we encourage the NLM and ISI to list both the group names and the names of individual authors, when both are published in the bylines and/or as author footnotes, in all citations and indexes. To ensure appropriate responsibility and credit, and to help direct citations and make them more readily retrievable, JAMA will include the research group name in the title or subtitle of relevant articles, even though this may appear redundant with some bylines, until it is clear that articles with group names in the bylines are consistently captured and cited in bibliometric databases.

THE JOURNAL will continue to require the identification of named and accountable authors for every article. Authors may choose whether they prefer to list individuals, or the research group name, or both individuals and the group name in the byline of articles. Recognizing the increasing complexity and collaborative nature of clinical research, we hope the options available to authors of large research groups will allow for appropriate identification of individual and collective contributions, credit, and accountability.