Settling authorship while the work runs
Authorship on a study is settled over the life of the work rather than at its end: a team agrees the rules before the research or the writing begins, records who contributes what as the work runs, and checks the list again at submission and during peer review.
Authorship before the work begins
Good Publication Practice 2022 collects its authorship guidance in supplement section G. In its lead author's summary of the process for determining authorship there are two main times to consider, and the first comes before any authoring begins, ideally at the investigator meeting for a clinical trial, or at study initiation for any other kind of study. The questions to discuss openly then are whether a group listing will be used and which kind, what the authorship requirements are, whether a target journal has been chosen, how somebody interested in serving as a byline author goes about it, and what the author agreement will contain, so that a person who cannot meet those responsibilities can opt out early. People should also know from the beginning how members of a publication steering committee are selected, and whether a lead investigator already chairs one.
The byline is not the only way to credit the people behind a study. Byline authors meet all the relevant authorship criteria, and where the study team is large a few of them may serve as stewards of the work on the group's behalf. An author group lists a large number of people who all meet the authorship criteria. An investigator group lists the study's investigators without requiring them to meet those criteria, but their names are still indexed in PubMed and their part in the study stays discoverable, which often matters for promotion. A study group is broader again, covering the investigators and other people involved, who may or may not meet the criteria, so several statisticians or several medical writers can be credited that way. A study group often makes investigators more comfortable about getting credit in the publication.
In one publication planner's view, the best moment to start discussing authorship is when research planning begins, ideally in a face-to-face meeting, although during the COVID-19 pandemic a good web call could work well and, in her recent experience, was even simpler. What matters is that the authors are selected before the study finishes, because by then there is at least a good idea of which publications could come out of it and who is likely to be most involved in each.
GPP3 (2015) asks that the rights, roles, requirements and responsibilities of all contributors, authors and nonauthor contributors alike, be confirmed in writing, ideally at the start of the research and in all cases before publication preparation begins.
One freelance writer with ten years' earlier experience in the pharmaceutical industry distilled the available guidance into a six-page document, Good Practice Guidelines for Publication, that she gives to every client. In her view ghost authorship arises from a lack of transparency, and a company's publication policy sets a standard at project inception, involves all stakeholders, keeps ethics at the forefront of the process, and raises the level of professionalism as an added-value service. She urged every medical communications business, from a sole proprietorship to a large company, to develop and maintain one, and a simple briefing document can serve to educate stakeholders and foster transparency during the publication process.
In a survey sent in August 2007 to medical writers at medical communication companies and at developers and marketers of health care products, 70% said that for research articles the authors were not involved until after the data tables had been developed, and in 5% of cases not until a first draft of the manuscript had been prepared. Close to half (46%) said authors were involved earlier than five years before, 50% saw no change, and 4% said later. The responses suggest that companies are encouraging earlier author involvement in manuscript development, although some still prepare draft manuscripts before recruiting the author.
Who leads the authorship discussion
In the view of people involved in producing Good Publication Practice 2022, talking to an investigator who wants to be an author without meeting all the ICMJE criteria is the responsibility of neither the medical writer nor the client: it is usually the lead author's, who holds those conversations within the authoring team, and a consensus the team reaches is not one to argue with. What matters most is transparency from the start: an investigator group or a study group lets the whole discussion happen before anybody has seen the data, and where no such group exists, the lead author or principal investigator should talk up front to everyone who might be an author, including about why somebody has not been included. The writer should not be driving those conversations or flagging that somebody is unhappy, although pushing a client at the start of a project to confirm that everyone knows what is happening and agrees to it is a good thing for a writer to do.
In one freelance writer's view, determining authorship or mediating authorship disputes is not, or should not be, within a freelance writer's purview. Where she has nonetheless taken on that work, she has been crystal clear with coauthors and clients about the guidelines and regulations and, knowing the politics of companies and academia, has addressed the issue directly. The politics themselves belong to the client, or to the chief author, investigator or scientist, and the writer cannot take responsibility for them.
Points that reopen authorship
Even where the potential authors were identified at the very beginning, the authorship discussion continues through the project, especially as new people become involved: a medical writer is often not involved at the beginning of an ongoing study, a new statistician or a new specialist may join, and someone may fail to start or may drop out of the study group, which is normal in any clinical trial. What matters is that every change to authorship is written down as the planning of the study and the disclosure of its results develop.
During manuscript development, everyone meeting the first ICMJE criterion, a substantive intellectual contribution to the design, conduct or analysis, should ideally be included unless the resulting number is prohibitive, and there should be a clear rationale for who is invited to do what.
In the lead author's summary of Good Publication Practice 2022, two later points reopen the question. At submission, who may now qualify should be evaluated again: a statistician who ran or designed new analyses may have become an author, and so may the medical writer, if they have been called upon to do enough to qualify. The evaluation should take place again during peer review, where comment consolidation and new data requests can also change the authorship of the paper. On Good Publication Practice 2022 itself, the work of consolidating its review comments helped make a very strong case for one contributor to become an author. Its lead author had no idea when first invited to take part that she would become lead author, and that contributor describes her own role as having evolved, like the lead author's, over the period the manuscript was produced.
In one case, feedback on a draft made its reviewer a coauthor. An author invited to write a statistics tutorial for a journal planned to write it alone, but a colleague's feedback on his draft was so substantive that he asked whether the colleague would be willing to be identified as a coauthor.
Tracking and declaring contributions
A medical writer can guide a research team through five practices to navigate authorship conflict. Plan authorship from the start, before the writing or even the research begins: at an initial meeting with the team, or with an established committee for a larger study, discuss the ICMJE guidelines and define what counts as a contribution qualifying for authorship on this project, so that every potential author knows the expectations, and the consequences of not meeting them, from the beginning. Establish each author's responsibilities, accepting that contributions may change as the project evolves, perhaps in a modifiable authorship grid that could be informed by standard taxonomies of author contributions. Agree, before any dispute arises, how disputes will be resolved, for instance through a neutral third party such as an administrator or a researcher far removed from the work. Keep track of who contributed and how, discuss authorship at regular intervals or major milestones and whether adjustments are needed so that all contributors qualify under the ICMJE guidelines, and update the grid or tracker with each discussion. And summarise every discussion in an email or document that all potential authors review and agree to in writing, then declare the contributions, and where appropriate clarify the order of authors, in the manuscript on submission.
Frameworks for complex trials
The researchers behind a five-step authorship framework for disclosing contributors to industry-sponsored clinical trials hold that current authorship guidelines are often too rigid for complex multi-centre, multi-disciplinary trials. They recommend that the difficult question of authorship be discussed openly before publication writing begins, and that author contributions then be monitored throughout the development of the publication so that individuals who merit authorship are recognised.
The Medical Publishing Insights and Practices initiative, which brings together pharmaceutical company representatives and ISMPP, held in 2015 that the ICMJE guidelines on authorship are too broad to allow valid and consistent assignment of authorship of publications based on clinical trials. Its members worked with other stakeholders to develop the framework, to address perceived shortcomings in the application of the ICMJE and other guidelines.
The framework grew from seven case scenarios illustrating difficult authorship decisions, which became an online survey sent to four stakeholder groups: clinical investigators involved in industry-sponsored trials, journal editors, medical writers, and industry-paid publication professionals. The results were discussed in two roundtable meetings to identify themes, and the survey was completed by 498 people.
A majority of respondents thought trial site management and a considerable contribution to patient recruitment were sufficient grounds for authorship. A majority thought a statistician who contributed to data analysis and interpretation, but not to trial design or manuscript drafting, should be added as an author of the final manuscript. And a quarter of the journal editors and clinical investigators thought medical writers should be listed as authors. The first two positions are potentially valid under ICMJE, but only where the person also meets the remaining criteria, whereas a medical writer does not normally qualify. Several respondents named the tension between what the guidelines permit and what they considered fair.
The framework has five steps. Establish a working group responsible for the remaining four, whose members need not be authors and should not be guaranteed authorship. Identify which trial-related activities are to count as substantial. Track and record those activities. Assess the record and invite the people responsible to be authors. Then check that everyone invited fulfils all the ICMJE criteria, adding or removing authors with the consent of all of them and documenting any change and its reason for internal transparency. A publication agreement approved by all trial contributors is suggested to define the procedures in the second and fifth steps.
In one commentator's view, the framework has the potential to make authorship decisions more consistent and reduce disputes, but a person invited to be an author at step four and removed at step five is unlikely to be content about it, whatever agreement they signed.
Authorship in a large multicentre trial can also be attributed by a score. The HF-ACTION Authorship and Publication (HAP) score, developed for a randomised trial of exercise training as a supplement to standard care in heart failure, assigns points for investigators' participation in trial enrolment, follow-up and adherence and for participation in committees and other trial activity. It was designed to enhance the rate of dissemination, to recognise investigator contributions to the successful conduct of the trial, and to harness individual expertise in manuscript generation. Its authors admit that it is not without limitations, but suggest it may be useful as a starting point for authorship decisions in multisite trials.
Changes to the author list
Under GPP3, during the development of a publication it may become necessary to add or remove an author, for example when an author fails to provide a substantial contribution or to approve the final version, and all the authors should then agree to the change. After submission, adding or removing an author should be considered only in rare cases, such as when the work changes substantially in response to reviewers' comments. A change of affiliation alone is not a valid reason to remove an author who meets the authorship criteria. An author who changes affiliation before the work is published keeps the affiliation where the major part of the work was done, with the current affiliation and contact details in a footnote or the acknowledgments.
Should an author die after completing a major part of the work, meaning after fulfilling the first two ICMJE criteria, posthumous authorship can be considered if all the other authors agree, and GPP3 suggests seeking advice on correct attribution and process from the journal's instructions or editorial office as the first step. If the journal agrees but requires submission forms to be signed, a supervisor may be the most appropriate proxy for a sponsor-employed author or a contractor, and otherwise a family member or a person with power of attorney should be approached; in all cases the deceased author's family should be told of the intention and asked to consent to the listing or acknowledgment. Where a local-language congress requires the presenter of an encore abstract to be an author, a name can be added for that purpose with all the authors' permission, identified as presenting on behalf of the group in the abstract byline where possible and at least in the presentation.
An investigator may be listed as an author without having provided any input or responded to requests to review the manuscript, and practitioners at a 2007 conference session offered five practical solutions. A publication policy in place at manuscript inception provides neutral ground to refer to. Offer to telephone the author to discuss the manuscript, which leverages guilt and which authors typically respond to; if there is still no response, remove the person from the author list. Give a hard deadline, stating that the name will be removed if no response arrives by a named date, an ultimatum authors often respond to. Where a sponsor needs to be moved to intervene, media reports of ghostwriting can be used to make the case. And where someone who has genuinely not contributed is still on the list at submission, contact the journal editor.