How Good Publication Practice developed through its editions before 2022
Good Publication Practice, the guideline for publishing company-sponsored research, went through three editions before 2022: a first in 2003, a second in 2009 and a third in 2015. Each built on the one before, adding detail and topics that the earlier one had left out.
The first edition, 2003
Good Publication Practice began at a meeting of academics, journal editors and industry representatives that the Council of Biology Editors, now the Council of Science Editors, organised in 1998 to establish clear guidelines and standards for publications of industry-sponsored biomedical research. It was developed by pharmaceutical industry representatives with the help of some editors. A working group of the council drafted it in 1998, and after review by the authors' companies the draft was published in four journals in 2000; the final version was published in 2003 by three authors in their individual capacities. It took five years from that first meeting to publication in 2003, as Good publication practice for pharmaceutical companies, the first document to set standards for industry-based manuscripts.
The first edition gave guidance on the ethics of presenting and publishing clinical trial data, to ensure that clinical trials sponsored by pharmaceutical companies are published in a responsible and ethical manner. In particular it set out to help avoid publication bias and to clarify the relationship between pharmaceutical companies and academic investigators. Its guidance covered publication standards, disclosure of potential conflicts of interest, what constitutes unacceptable prior or concurrent publication, identification of clinical trials, authorship and the proper role of professional medical writers. Besides publication bias and the relationship between sponsors and investigators, the issue it primarily addressed was redundant publication of data. Its scope was restricted to a relatively small set of issues: the obligation to publish everything, the role of professional medical writers in assisting with manuscripts, and a first brief approach to authorship. Medical writers rapidly took it up as guidance within their companies, but even at its inception it was evident that many topics had not been considered and that a more comprehensive document was needed.
The second edition, 2009
The second edition, GPP2, titled Good Publication Practice for Communicating Company Sponsored Medical Research, was published in 2009, six years after the first. It was released in April 2009 at the annual meeting of the International Society for Medical Publication Professionals (ISMPP). ISMPP spearheaded it, and ISMPP members wrote it. It was published in the BMJ. The rationales for the update were the publication of new guidelines and reports from other organisations, changes in the regulatory environment, increased media coverage of the issues, a need for expanded scope, particularly on authorship, reimbursement, the role of medical writers and publication planning, and the inclusion of medical device and biotechnology companies.
GPP2 was written by a larger author panel than the first edition's three authors: twelve, representing pharmaceutical companies, publishers, communication agencies and independent medical writers. Before submission, a wide review panel that included academics and journal editors reviewed the guidelines. A fourteen-member ISMPP steering committee revised and expanded the first edition, then sent it to 193 consultants from all areas of medical publishing, AMWA representatives among them, whose 116 blinded sets of comments the committee evaluated before the final version was published in a peer-reviewed journal.
Like the first edition, GPP2 is designed for companies that sponsor clinical trials and for any company or individual working on industry-sponsored publications, freelance writers, contract research organisations and medical publications and communications companies included. It applies to oral and audiovisual presentations at scientific meetings as well as to journal publications, peer-reviewed or not.
GPP2 was a more complete and more comprehensive document than the first edition. It was a refinement of the positions the first edition stated. Its new sections address recent developments such as clinical trial registration and the ethics of results disclosure. It introduced or supported written publication agreements, publication steering committees, checklists and the contributorship model, while reinforcing the authorship guidelines of the International Committee of Medical Journal Editors (ICMJE) and adherence to trial registration and results-posting requirements. It cites a number of key references, among them AMWA's position statement on the contributions of medical writers to scientific publications, and makes key recommendations on authorship, transparency and disclosure, and publication planning practices that may benefit many who work in medical writing.
The key differences between the second edition and the first include these, two of them on matters the first edition gave no guidance on:
- Authorship. As in the first edition, follow the ICMJE's criteria where possible, or list contributors where the journal requires it, and apply whatever criteria are used in the same way to external investigators and to company employees. In addition, before writing begins, one author, a lead author who may also be the guarantor, should take the lead for writing and managing each publication or presentation, and one author, identified as guarantor, should take overall responsibility for the integrity of a study and its report.
- Acknowledgment of professional medical writers. Where the first edition said that the medical writer should be acknowledged, the second gives detailed guidance on how contributors, sponsor companies included, should be acknowledged, and recommends that authors and the professional writers working with them use a published checklist to discourage ghostwriting and take particular care to acknowledge writers' contributions appropriately and describe their funding.
- Reimbursement. Authors should not receive honoraria for peer-reviewed articles or presentations, while reimbursement of travel expenses and payment for specialised services such as statistical analysis may be reasonable.
- Documentation. Companies should develop policies on the types of documentation to maintain and for how long, for example retaining main versions of the draft to document how comments on previous versions were incorporated.
A member of the committee that developed GPP2 named these as its most important new recommendations:
- Incorporate the guideline's requirements into a written agreement between the sponsor and the author, describing the roles and responsibilities of each.
- Form a publication steering committee within the company to oversee publications.
- Give authors and other contributors full access to the sponsor's data.
- Pay nobody, honoraria included, for authorship of articles or presentations.
- Generate publication plans.
- Comply with the ICMJE's uniform requirements for authorship, with guidance given for where a journal's criteria deviate from them.
- Use the acknowledgments section for clear and concise descriptions of the roles of each author or contributor, the medical writer included.
- Include the clinical trial identifier and state whether the publication or presentation is the primary one for the trial's results, with definitions of primary and secondary publication provided.
- Comply with established reporting standards.
- Provide clear methodology for systematic and comprehensive review articles and meta-analyses.
- Maintain documentation of how a publication or presentation is initiated, conducted and finalised, with recommendations for what to archive.
GPP2 notes that it may be useful to form a publications steering committee of members of the study steering committee and the protocol development team, other investigators in the clinical programme and sponsor employees.
GPP2 recommends that authors follow established reporting standards such as CONSORT for randomised controlled trials and PRISMA for systematic reviews and meta-analyses.
In the view of one of its developers, following GPP2 demonstrates integrity, completeness, transparency, accountability and responsibility. Its recommendations also include specific language to use and conflict-of-interest disclosures.
As a more detailed document with many concepts more precisely covered, GPP2 was ideal for helping publication professionals establish internal guidelines and ways of working. Having such externally written guidance also lent credence to the idea that publication professionals brought value to their employers, and reinforced writers' ability to insist on ethical working practices. AMWA's Board of Directors endorsed GPP2 in April 2010. The Board's vote to endorse it was unanimous.
The third edition, 2015
As with the first edition, GPP2 left numerous unanswered questions and requests for more detail and additional topics, so the third edition, GPP3, continued to build on the earlier guidelines. ISMPP was leading its development in 2013. GPP3 was the work of ISMPP members, and an article describing it was submitted to the Annals of Internal Medicine in January 2015. Sponsored by ISMPP, it was published there in September 2015. Peer reviewers were expected to have some influence on the final content of the article. GPP3 reflects the changes in legislation on the dissemination of data from clinical trials, mainly focused on the pharmaceutical industry, and is intended to serve as a basis for publication professionals to establish transparent and ethical working practices within that industry. It extended the earlier versions' focus on the transparency and integrity of medical publications and set out to broaden adoption of good publication practice globally; by 2017 a Chinese translation was available and a Japanese one was under way.
GPP3 focuses on the core values of the first two editions and adds further detail and an improved organisation to enhance clarity and eliminate redundancy, and it updates the sections already in GPP2 for how publication and data dissemination practices had evolved. New in it were sections on core publication principles and on data sharing. It opens with a list of ten key publication principles, which are intended to support the guideline's six core principles of integrity, transparency, completeness, accuracy, accountability and responsibility and can also serve as a checklist for authors and medical writers. It states explicitly what integrity, completeness and transparency mean, who is accountable for the results reported, and what the responsibilities are of the authors, medical writers, statisticians, data managers and everyone else involved in a clinical trial that has to be disclosed after it ends.
GPP3 addresses authorship, on which the ICMJE had revised its requirements in 2010 and 2013, adding a fourth criterion. It sets out to identify and define authorship clearly, and specifically to eliminate guest authorship and ghostwriting once and for all. Publication professionals were aware of those practices and avoided them, but having them defined in the guideline makes that knowledge easier to communicate to colleagues and ensures the practices are recognised as unacceptable. The edition gives additional guidance on payment and reimbursement of an author's time for their role in developing a manuscript, an issue on which even its authors had differing perspectives, and addresses author number, author order, deceased authors and authors no longer with the company in a table.
GPP3 supports the contributorship model of authorship, in which each author's role is clearly defined and potential conflicts of interest, financial or otherwise, are disclosed even where the target journal does not ask for them. It recommends calling that information Disclosures, a term that carries no negative connotation and is more likely to encourage disclosure of both financial and non-financial interests. The model may also reveal gaps in contributions the author panel should cover: if nobody is identified as having performed the statistical analyses, it is unclear who was responsible for the accuracy of the data and the appropriateness of the analytical methods. The edition also discusses author groups for large trials.
GPP3 has a section on the role of medical writers and what is expected of them. It addresses two common questions for publication professionals, how they justify their role in preparing manuscripts and why they are not then authors, notably by presenting published data on the importance of medical writers in improving the quality of submitted manuscripts. It recommends how writers should work with authors, the key recommendation being to establish defined roles and responsibilities before writing starts, and discusses how a writer's role may in certain circumstances result in authorship. As GPP2 did, it recommends written agreements for authors, medical writers and agencies and the establishment of publication steering committees, and it recommends that, for transparency, any writing contribution be acknowledged along with the source of its funding.
Since GPP2, an important area of evolution had been the dissemination of clinical trial data after trial registration, through data posting and data sharing with researchers. GPP3 recognises and fully endorses all aspects of clinical trial reporting, while holding that reporting does not substitute for publication as a means of presenting and explaining the data in context. Data posting does not constitute prior publication, nor does it discharge the ethical obligation to publish, because the researcher has to present the work in the context of current knowledge and note the contribution it makes.
Under GPP3, any publication of clinical trial data, whether abstract, poster or paper, should carry the appropriate trial registration identifier so that readers can identify the study, which also helps ensure that data are not published in duplicate, even unintentionally. Exceptions to duplicate publication can be made for encore presentations of abstracts and posters at congresses in different specialties or geographies. Every effort should be made to minimise plagiarism, including self-plagiarism, which the guideline addressed for the first time in this edition. The primary publication must appear before secondary articles, which must clearly identify and refer to it. The edition gives some guidance on timing, noting the effect data posting may have on publications, and touches briefly on the types of scientific article, with some indication of the guidance already available for different types, such as PRISMA for systematic reviews. That section also covers encore presentations at different congresses.
GPP3 was endorsed by an increasingly wide range of organisations concerned with scientific communication and medical writing, including EMWA, AMWA, the Committee on Publication Ethics, the European Association of Science Editors and the Japan Medical and Scientific Communicators Association. In 2016 GPP3 was the latest evolution of an established process, and a fourth edition was expected to take account of further evolution in data dissemination and publication, its timing and content yet to be determined. Typical comments from reviewers and users of GPP3 asked for it to be made for academics too and sent to universities, and suggested that a next step might be guidelines for non-company-sponsored medical research. With the support of the organisations endorsing GPP3, the fourth edition might become the model for all writers of medical communications, not only of industry-sponsored studies, for increasing transparency and the quality of clinical trial reporting. In the meantime the GPP3 steering committee, with ISMPP, set up an online repository of questions and answers and resources on publication practice, with further questions accepted by email to the authors, and the site might in future also serve as a platform for recruiting new members to the GPP team to work on the fourth edition.