Deciding who qualifies as an author
Who qualifies as an author turns on the criteria a contributor has to meet, on how journals and conferences apply them, on what they mean for other contributors and for a professional writer, and on how each contributor's part in the work is recorded.
The growing number of authors
Papers indexed in MEDLINE/PubMed carried an average of 1.50 authors in the 1950s and 5.48 in 2015, according to the US National Library of Medicine's statistical reports, while the number of collective author names, also called group or corporate names, did not increase over the same period.
Until the 1950s publications were largely written by one author. Between 1945 and 1988 medical publications gained 1.26 authors every 15 years, against 0.41 in other branches of science, and across the sciences the rise in authors has paralleled the rise of international and interdisciplinary collaboration. The rise leaves research teams two decisions: who qualifies for authorship, and in what order to list the authors. Both can be harder among junior researchers not versed in authorship practice and in teams with unethical authorship practices. The tension they can cause in a research team may also relate to power dynamics and work culture. To overcome these challenges, research teams may look to medical writers and editors, inside or outside the team, for guidance. So medical writers and editors need to understand the authorship guidelines fully and be prepared to share best practices for deciding who qualifies for authorship and who should be listed in the acknowledgements.
The earlier three-condition criteria and their critics
The earlier form of the criteria of the International Committee of Medical Journal Editors (ICMJE) had three conditions, all of which had to be met: substantial contributions to conception and design, or acquisition of data, or analysis and interpretation of data; drafting the article or revising it critically for important intellectual content; and final approval of the version to be published. In 2008 it was the definition of authorship most used in the biomedical sciences.
The earliest expression of those criteria came from Richard M. Hewitt, an AMWA member, whose Medical Writer's Ten Commandments of 1957 told writers not to allow their name to appear as a co-author without some authoritative knowledge of the subject, a part in the underlying investigation, and labour on the report to the extent of weighing every word and quantity in it.
One lecturer in ethics argued that the three-condition criteria were unethical and should be changed. In a hypothetical study, three researchers contribute in different ways to its design, its implementation and its reporting, and none of them meets all three conditions. In his view, having a great idea, sharing it with colleagues and approving what they do with it is to co-write a paper, as gathering and analysing data is, and as redrafting and reviewing a paper is. So, he argued, meeting one of the three should be sufficient for legitimate authorship.
Another critic argued in 2013 that the three-condition criteria were vulnerable to misuse, and proposed a further criterion: free and unfettered access to all raw data.
One analysis of retracted articles holds that the three-condition criteria did not address the integrity of the data underlying a manuscript, and that the ICMJE acknowledged the shortcoming by advising that each author should have participated sufficiently in the work to take public responsibility for appropriate portions of the content. There is no consensus about the responsibility of coauthors to maintain data integrity. No one is in a better position to detect data fabrication or falsification than a coauthor. Most scientists found guilty of misconduct by the Office of Research Integrity engaged in falsification, fabrication or misrepresentation of data, and coauthors should be able to detect and prevent such manipulation. So, the analysis concludes, coauthors should take responsibility for the integrity of the study data as well as for the usual authorial tasks. Coauthors must know the data and press the first author to describe them accurately.
The four criteria and the accountability they ask of authors
Since the ICMJE's revision of August 2013, an author must meet all four of its criteria:
- Substantial contributions to the conception or design of the work, or to the acquisition, analysis or interpretation of data for the work.
- Drafting the work or revising it critically for important intellectual content.
- Final approval of the version to be published.
- Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Under the ICMJE recommendations, everyone who meets the first criterion should have the opportunity to take part in the review, drafting and final approval of the manuscript. So someone who qualifies under it cannot be left off the author list simply because they were denied the chance to meet the second and third.
The fourth criterion was the new part of the 2013 revision. It was added because of cases of author misconduct in which authors denied responsibility, and the committee's commentary goes further: each author of a paper needs to understand the full scope of the work, know which co-authors are responsible for specific contributions, and have confidence in the co-authors' ability and integrity. One medical writer expected in 2013 that the fourth criterion would be the most difficult for authors to comply with, and already found it difficult to get most contributors to comply with the first three.
A former editor of the BMJ holds that each author must assume full intellectual accountability, and gives a practical test: an author who could not take the place of a lead author taken ill before a conference presentation, give the presentation and answer all the questions is not intellectually accountable. In his view about half of authors do not meet that standard.
The fourth criterion makes all authors accountable for the work as a whole, not only for their own contributions, so authorship indicates responsibility for the integrity of the work as well as credit for it. Disagreements among the authors about the data or the opinions in the manuscript must be resolved before submission, and all authors have a responsibility to resolve questions raised after publication about the accuracy or integrity of the work. In 2014 many medical writers might not yet have been familiar with the criterion.
Writers need to communicate that responsibility clearly to the authors of the articles they prepare. Many peer-reviewed journals have adopted the ICMJE criteria into their publication policies, so it is essential that writers discuss with authors the importance of understanding and following the ICMJE recommendations and other ethical guidelines, to maintain transparency, to protect against accusations of ghostwriting and to ensure compliance with journal policies. Many authors may be unaware of the ethical issues surrounding authorship, sponsorship and potential conflicts of interest, or of the controversies, recent in 2014, over ghostwriting and guest authorship of sponsored clinical study findings, so writers have a responsibility to inform them on those issues and to take a strong ethical stance in the face of resistance from either party. Informing clients about content and ethics guidelines such as the ICMJE's is part of the professional medical writer's responsibility to maintain the highest ethical standards.
The venue's rules and the number of authors
Under Good Publication Practice 2022, what constitutes authorship follows the rules of the journal or conference a team is submitting to. The ICMJE itself says that journal editors and conference organisers are empowered to make those decisions, so a venue whose criteria differ from the ICMJE's is within its rights. The 2022 edition also removed the limit of twelve authors that the earlier edition had set. Reviewers' comments on the change were roughly evenly divided between welcoming it, since teams often need fourteen authors, and fearing that ninety people would want to be listed. So the final text asks for a reasonable number of authors, sets no limit per function and follows the ICMJE position that no specific limit should be set.
Contributors acknowledged rather than listed
Many journals give authorship guidance on their websites, most often recommending the ICMJE criteria, and many ask for a description of what each author contributed and often publish it with the paper. A contributor who does not meet all four criteria should be acknowledged instead. Contributors who supplied funding, research materials, lab management, administrative support, or writing and editing assistance may not meet all four, and each should be thanked for the specific contribution, such as providing funding, critically reviewing the manuscript or collecting data.
The list of authors should be restricted to those who made a major contribution to the study, with minor contributors acknowledged at the end of the paper. Acquiring funding, collecting data, and general or administrative supervision of the research group do not by themselves make someone an author. Nor does being a department head who is aware of the study but takes no active role in the investigation or in the preparation of the report.
Honorary and ghost authorship
Medical writers and editors need to be prepared to advise research teams against unethical authorship practices, including honorary and ghost authorship. Honorary authorship is unethical and can be considered research misconduct, so medical writers and authors need to be prepared to advise against it. A professional writer who qualifies under the ICMJE criteria should be on the author list. Anyone who wrote a work or otherwise contributed substantially should put their name on it, and anyone who did not should not allow themselves to be listed as an author. No one should allow themselves to be listed as an author, especially a lead author, on a paper written by someone in a pharmaceutical company when they had no hand in the writing. Several of the 13 coauthors of a 2005 paper in The Lancet by Sudbo and colleagues claimed that they were not aware of the submission or the full result.
Omitting significant contributors from the author list is a form of scientific misconduct, and conferring authorship on people who have not made substantial contributions is equally questionable. That practice is very common among group leaders and student supervisors who take advantage of inexperienced junior researchers, and abusive co-authorship of this kind has been called the 'white bull effect'.
Local researchers from low-income countries
As of 2024, recent ICMJE recommendations state that editors should be aware of the practice of excluding local researchers from low-income countries from authorship, and that including those researchers adds fairness and adds context and implications to the research.
Authorship tasks and presentation tasks
A pilot survey of experienced medical writers and editors, published in 2008, rested on its author's premise that a scientific document has two broad aspects, content and presentation. In that premise, creating, selecting or modifying content are authorship tasks: analysing data, reviewing the literature, drawing conclusions, and abstracting facts and arguments from other sources. Organising, clarifying, explaining and condensing that content in words, tables or figures are the presentation tasks of writing and editing, the packaging of insight or information that authors created. By the survey's definitions an author creates and selects content and may also write or edit, while a writer or editor deals solely with presentation, and what makes an author an author is the tasks performed, not the person's job title, academic preparation or position in the organisation.
Sixteen experienced writers and editors took part in the survey, which set seven scenarios. In its author's reading only three involve authorship: writing a review article on a new drug's adverse events from the client's talking points, having searched the literature and selected both the articles cited and the information drawn from them; finding all published cases of a rare anomaly and writing the literature review for another author's case report; and writing the results and discussion sections of a clinical study report from the raw data, where few if any reviewers will go back to the data to see what was missed. The respondents agreed on all three, and most of them also called two other scenarios authorship. Thirteen of sixteen said so of drafting a chapter on hip pain, to an outline the client approves, from thirty articles the client supplied, most with sections highlighted, and polishing the final version after the client has reviewed and revised the draft. Eleven of sixteen said so of drafting a paper for a principal investigator from interviews with him, data printouts and access to the statistician, where the writer had attended the department's meetings on the research and helped write the original grant, and the investigator and his group would review and guide the work. Four thought, as the survey's author did not, that completely reorganising and rewriting a poorly written paper by a non-native English speaker was authorship. The group was almost evenly divided on the scenario in which the writer, preparing a poster from an accepted abstract, finds that the data and analysis do not support the abstract's conclusions and the client accepts the writer's reasoning and revises the poster.
For each scenario the survey asked two separate questions, whether the writer was being asked to do authorship tasks and whether the writer should be named as an author. Mixed responses, a yes to the first with a no to the second, were given for every scenario. Some respondents said that the degree of physician oversight was not clear in the chapter scenario, which may explain its five mixed responses. In the author's reading the writer in the poster scenario is acting as a peer reviewer, raising a question the authors are free to act on or not, and that contribution is not authorship. The scenario was included because no one had raised whether peer reviewers' contributions should be acknowledged or their names appear on the article.
Whether a writer should be an author
On manuscripts, medical writers often go well beyond basic services. They are often the main force behind an article and may provide much more intellectual input than putting the client's ideas into words, including interpretation and sometimes analysis of data, production of graphs and tables, and input into how the data and ideas are presented, often more than some of the listed co-authors. So many writers feel that they should qualify for authorship, perhaps even first authorship.
In early 2013 the ICMJE's guidelines were the main ones for determining who should receive authorship, and Good Publication Practice 2 referred to them on authorship. Those guidelines considered an author generally to be someone who has made substantive intellectual contributions to a published study, held that biomedical authorship continues to have important academic, social and financial implications, and stated that an author must take responsibility for at least one component of the work, should be able to identify who is responsible for each other component, and should ideally be confident in the co-authors' ability and integrity. A writer who makes the contributions above, and who is obviously responsible for drafting the article, probably cannot meet the last two of those statements, nor the third condition of the three-condition criteria. In effect, authorship implies accepting liability for what is published, which is very likely something a writer does not want and could mean buying expensive liability insurance. So writers should not be listed as authors unless they are willing to take public responsibility, and therefore accept liability, for the article and its contents.
EMWA's guidelines on the role of medical writers say that writers should not agree to be listed as authors where they do not meet the target journal's authorship criteria. To qualify under the ICMJE criteria, a writer would need to have made a substantial contribution to the analysis or interpretation of the data and to feel able to take public responsibility for the research, so in practice professional writers are unlikely to be named as authors on primary research publications. Where a journal supplements the by-line with a contributor list stating each individual's contribution, it might be appropriate to list a writer who prepared a first draft or made another significant contribution, following that journal's requirements.
The analysis of retracted articles holds that, because coauthors should take responsibility for the integrity of the study data, a medical writer who helps in the publication process but has no insight into data integrity should not be listed as a coauthor.
Under the four criteria, a professional writer who helps authors prepare a paper usually does not qualify, and one who does is a rare exception. A writer involved only in the writing, and not in the design, development, data collection, analysis or interpretation, does not meet the first criterion.
Two experienced publication writers have seen such exceptions. In one, something like a patient-preference study, the writer had to collate and analyse a whole body of responses. In another, a writer who had seen the results of the statistical analysis proposed exploring a question further, the new analysis gave new insight into the study's outcomes, and the writer was asked onto the author list, which satisfies the first criterion. A writer may also qualify as an author of a review article, for example after an extensive literature search. One of the two writers had seen this once or a couple of times, where the writer had worked out which publications went into the review and driven its structure.
The record of who did the work
The discovery of streptomycin, made in the early 1940s, led to a major dispute between a professor and his doctoral student. The student, working under the professor's guidance, performed the experiments that, after ten unsuccessful ones, isolated two strains of a soil organism producing the antibiotic. The initial manuscripts and the patent were in both his and his supervisor's names, but the 1952 Nobel Prize went to the supervisor alone. According to a book on the dispute, the supervisor later rewrote the story in a way that played down or ignored the student's contribution while the institution promoted him, and the student sued, in a case settled out of court. The student struggled to get jobs and drifted into obscurity, until a British microbiologist tracked him down in the late 1980s and interviewed him in 1989, and he made an emotional return to the institution, which belatedly honoured him for his work.
In the book's reading, the passive voice of the research papers made it almost impossible to determine who had done what, statements of how each author contributed to a study are a relatively new thing, and with nothing else to go on the Nobel Committee may have placed great weight on the supervisor's name being on all the key papers and the student's not. Doctoral students often have little power in such a situation because they fear that complaining will upset their supervisor and earn them a bad reference that could damage or end a scientific career, and the book gives that fear as the reason the student signed over his patent rights. In that period, scientists who directed laboratory programmes of any significance regularly appeared as senior authors on all papers from their laboratories, and in the experience of one reviewer of the book, writing in 2012, the practice still went on.
Contributorship
As of 2012 the journal Neurology had its own authorship policy, which departed from the ICMJE criteria toward a contributorship model, identifying everyone who contributed to the study, wrote the report, paid for the research and so on. Its criteria for authorship were design or conceptualisation of the study, or analysis or interpretation of the data, or drafting or revising the manuscript, and four requirements went with them: all authors acknowledge all versions; those who do not qualify as authors should be listed as co-investigators or contributors; any paid medical writer who wrote the first draft or responded to the reviewers' comments must be included in the author byline; and all authors complete and sign authorship forms recording roles and contributions, disclosure forms listing all sources of potential bias, and copyright transfer agreements.
The journal hopes that the policy will foster greater transparency and disclosure and help avoid honorary and ghost authorship. Two of its editors hold that including professional writers as authors is transparent, fair and anti-discriminatory, credit being given where credit is due, and that research is becoming more complex and collaborative, which increases the challenges of transparency in authorship and disclosure. They offered the policy as a starting point in the effort to improve transparency and suggested that more journals adopt the contributorship approach in their instructions for authors.
The proposal to standardise contribution reporting
In February 2017 the president of the National Academy of Sciences convened a group from leading journals and scientific organisations at a retreat to discuss how to promote standards that would increase transparency in authors' contributions to research papers, and the outcome was posted as a preprint on 20 May 2017, open for comment. Its proposed adaptation of the ICMJE statement expects each author to have made substantial contributions to the conception or design of the work, or to the acquisition, analysis or interpretation of data, or to the creation of new software used in the work, or to have drafted the work or substantively revised it; to have approved the submitted version and any substantially modified version that involves the author's contribution; and to agree to be personally accountable for their own contributions and for ensuring that questions about the accuracy or integrity of any part of the work, even parts in which they were not personally involved, are appropriately investigated, resolved and documented in the literature.
The proposal's other recommendations are addressed to journals, research institutions, funding agencies and scientific societies. Journals should adopt common and transparent standards for authorship, outline the responsibilities of corresponding authors, adopt the Contributor Roles Taxonomy (CRediT) for attributing contributions, include that information in article metadata, and encourage authors to use the digital persistent identifier ORCID. Research institutions should hold regular open conversations on authorship criteria and ethics. Funding agencies should adopt ORCID and accept CRediT. Scientific societies should promote authorship transparency by implementing these recommendations through their meetings and publications programmes. The taxonomy itself, implemented by a few journals by then, defines these contribution roles: conceptualisation, methodology, software, validation, formal analysis, investigation, resources, data curation, writing and original draft preparation, writing and review and editing, visualisation, supervision, project administration, and funding acquisition.