Good Publications Practice in the Workplace
GPP 2022, Professional Engagement, and Artificial Intelligence in Publications Practice
Lisa DeTora
Hofstra University
Meera Kodukulla
Jennie G. Jacobson
Jacobson Medical Writing, Inc.

he current iteration of Good Publication Practice (GPP) Guidelines for Company-Sponsored Research was published in Annals of Internal Medicine in late summer 2022. These guidelines included important new topics previously reviewed in Global Forum, such as working with patients and helping to establish scientific and medical publications as a key function not just within clinical research but across the scientific endeavor of biomedical research in company-sponsored settings.

Lead author on this GPP update Lisa DeTora discusses aspects and practices of authorship and social media with industry publication colleagues Meera Kodukulla, a volunteer reviewer of the GPP 2022 update and senior author of the ISMPP Authorship Recommendations; and Jennie Jacobson, president-elect of the Delaware Valley Chapter of the American Medical Writers Association. They discuss GPP, including author selection based on the International Committee of Medical Journal Editors (ICMJE) authorship criteria—specifically its first criterion, which refers to intellectual contributions to study design, conduct, and data analysis, and interpretation (hereafter referred to as “criterion one”).

Lisa DeTora (LDT): Authorship is a key issue in publications management because it has implications for ethical conduct, professional relationships, and research integrity. Developments like GPP 2022 or the International Society for Medical Publication Professionals (ISMPP) authorship recommendations and its algorithm reinforce what was already understood as good practice. In your opinion, what are key practical issues facing the profession now?

Meera Kodukulla (MK): In the context of authorship, the key challenge remains defining authorship eligibility, particularly with regards to the ICMJE criterion one. Contributions frequently are not identified and documented during research or study conduct, and this leads to a lack of universal interpretation and application of ICMJE’s first authorship criterion (or criterion one) not only across sponsor organizations but within sponsor companies. The ISMPP Authorship Task Force published recommendations providing further clarity to interpreting ICMJE criterion one, the first time that a group of expert publication professionals specifically parsed out ICMJE criterion one. While the ICMJE criteria provide overarching recommendations, there is still a lack of clarity around how to interpret criterion one. The ISMPP authorship recommendations include definitions of substantial contributions intended to support both publication and authoring teams in early and transparent identification and documentation of contributions.

LDT: I understand that the authorship algorithm is a voluntary tool that can be accessed by ISMPP members. Can you explain the authorship algorithm for those who are not already familiar with it?

MK: The core goal of the algorithm is to help document and rank contributions in the authorship byline. The authorship algorithm is essentially quantifying relative substantial contributions to each of the four ICMJE criteria for a study: Concept and design, acquisition of data, analysis of data, and then interpretation of data. There is scoring and there is evaluating, and this results in a set of total individual scores for each contributor.

This gets us to a universal weighting scheme, which is then evaluated across the board in three different categories based on the scoring: significant, moderately significant, or minimally important.

I want to add that while we have the ISMPP authorship algorithm, many sponsor organizations have their own algorithms that they continue to use. Whether you use your company-specified algorithm or adopt the ISMPP authorship algorithm, the methodology is the same and it helps us arrive at an authorship byline that accurately reflects the contributions of all authors.

LDT: Jennie, would you share the key impacts of the updated GPP guideline on your day-to-day work?

JJ: Most companies and agencies were already GPP compliant, so the changes have been fairly minimal. Which is not to minimize the importance of GPP 2022, but it was not a sea change for industry, especially in terms of clinical research publications.

MK: There is increased clarity in GPP 2022 around specific topics, and most large sponsors and even smaller companies have adopted the GPP recommendations into daily practice across the publication development process. This is a step up. This has contributed to publication professionals within pharma and sponsor organizations taking the GPP 2022 recommendations alongside the ISMPP authorship recommendations and the ICMJE criteria to re-educate stakeholders, whether they work in research, in real-world evidence, or in medical affairs. Taken together, these resources offer a supportive tool for understanding and adopting not just the authorship aspects but the GPP 2022 recommendations as a whole.

LDT: One of the things we wanted to do (and did do) while designing GPP 2022 was break the supplement into sections so that people could use them for training. Training is an important element of the leadership part of the medical publication professional role.

JJ: The improved organization has been very helpful, and I also appreciate the increased emphasis on the patient and diversity brought out in this version.

LDT: It was the first time that GPP included patients or talked about diversity and inclusivity. We also expanded the scientific scope, hopefully taking what people already knew about doing clinical trial publications well and broadening that knowledge out across organizations and into other scientific areas.

Recent legal advice regarding social media posting suggests that regulators do not view social media as scientific exchange, so it’s not included in the regulatory “safe harbor” with peer-reviewed publications. Do you have any comments on social media or how you handle it?

JJ: In general, I think you shouldn’t handle it. Most of my clients have an office of social media, and other professionals should leave that entirely to them and to the legal team because they’re experts in what can and cannot be done. Different companies interpret things differently, so what one company will see as acceptable, another will not. If you’re working in only one company, you may be more familiar with their guidelines, but in general, I would leave it to the experts. The role of the publication professionals here is simply to let those social media experts know, “We have a publication coming out. Do you want to do anything about it?” You definitely want to be communicating with them to see if anything should be happening on social media, but you should not be doing it yourself.

MK: That’s exactly right, Jennie, and this applies across the industry. We actively partner with our corporate affairs or corporate communications or scientific communications colleagues, who lead decision-making related to social media postings. Publication professionals within a company are not directly involved in putting out any social media postings.

LDT: It may or may not be the role of the publication professional to intervene, but if one of your investigators is known to be associated with your company, it might be nice to caution them about tweeting or posting about any products because that could also be viewed as promotional. In other words, even if you are not handling social media, investigators may not be as cautious as company professionals. Academic authors also may not know the limits of the scientific safe harbor regarding specific products—there is regulatory guidance about where and how off-label scientific exchange can occur in a “safe harbor” without being considered promotional. That knowledge will depend on their prior experience, and it’s often up to publication professionals or medical communication experts to explain.

Our next hot topic is artificial intelligence, or AI. I know that there’s a role for AI in regulatory documentation. Some companies have built their own AIs to write narratives, but do we have any predictions for whether this type of methodology will be moving into publications, which is a far different field?

JJ: I think AI has already moved into publications, to some extent. No one, I hope, is using it to write. You cannot put your data in it because it then becomes public domain unless a company has its own private AI toolkit, and I am not aware of any company that has such a toolkit for publications. You obviously do not want your data in the public domain. People are using AI to provide summaries of manuscripts; anybody who does that should then read the manuscript to double-check for correctness and completeness of any points you’re using from that summary.

People have used AI in their research, too. Often, it’s much easier to edit something than to start with a blank page, so I’ve heard of people putting the topic of their introduction or their manuscript into AI to ask, “What are the three most important points about this?” and using that as a starting place. But absolutely no one should be putting data in it. Nobody should be relying on it. We know about hallucinations and other inaccuracies. Frankly, I don’t think AI writes nearly as well, as far as sentence flow and other things, as a professional medical writer.

LDT: Chat GPT is very journalistic in its style, which is not always appropriate for biomedical publications.

MK: There’s still lots more to come in AI. We’re barely scraping the surface. There may be elements of AI that we may be able to embrace into publication development per se, but there could also be potential areas of risks. When I think about the end-to-end publication development process: Yes, we can have AI put together a shell or preliminary draft, but there is so much of the human element when it comes to interacting with our stakeholders and working with healthcare practitioners. There’s accuracy and credibility and so much more at stake when you look at the final publication output.

We also have to ask, “What is the journal side of things? What do journals believe?” As of now, we are required to disclose if an AI tool was used in the development of a publication. This area is still new. There’s so much more to come, and I’m looking forward to seeing where this will end up in the publication space. And, as Lisa said, (regulatory) medical writing is an altogether different aspect.

LDT: This is such an interesting topic. There’s the kind of artificial intelligence that you can use to read a script, instead of hiring a voice actor, if you’re making a video or a podcast. Or you can use AI for a lay summary of something; it can take text with more complicated technical language and translate it for other users. But as Jennie pointed out, AI-generated text needs to be double-checked because it’s just an algorithm. All it can do is put together the most statistically probable combination of words, which may or may not actually be accurate in terms of human comprehension and scientific integrity. Remember that it’s not really an intelligence, it’s just an algorithm.

JJ: If you’re using AI to generate a plain language summary, it had better be a plain language summary of data that’s already in the public domain. Because if you’re using undisclosed data, you’re putting it in the public domain.

LDT: Good point. What are our predictions about the increased emphasis on inclusivity of patients or other groups in the publications profession?

JJ: I do see inclusivity as increasing, although not nearly at the rate that we would all like it to increase. The diversity of patients in the trial is outside the influence of publications except for us to cheer on our colleagues on the clinical side. They need new and different recruiting strategies. But on the publications side, if we can get large enough study populations, doing subgroup analyses to see if drugs work as well across different races and ethnicities can be useful. That’s something we should be encouraging people to look into.

MK: When it comes to patient involvement in publications, I see more of this happening at the protocol development stage, seeking patient input. Once they’re involved, you’re talking about contribution to criterion one, essentially contributing to the study design, and then you can pull them through when looking at the overall contributions to the research or study. They definitely can be included as authors. There are certain therapy areas where this might be more predominant and in use than others.

LDT: These are excellent points. In chronic conditions, rare diseases, there’s much more patient involvement as you’re building the studies. More and more medicines agency initiatives want to include patients in design of the trial and its endpoints. This kind of practice builds in that type of intellectual input early, and then builds in the opportunity for including patients as authors later on.

JJ: I had an opportunity to be involved in an interesting paper about the patient experience of their disease, written by a patient and her physician. In the past, that would have just been written by the physician. But for a rare disease, a manuscript like that can be a very useful type of resource for patients and physicians.

LDT: The Multi-Regional Clinical Trials Center of Brigham and Women’s Hospital and Harvard does a lot of work with patient advocacy groups and patient inclusion in various children’s initiatives for pediatric studies. If people are interested in that from a more academic perspective, that’s certainly a rich source of information.

The ICMJE noted, and we echoed in GPP 2022, that they’re looking for more inclusivity from all regions where research is conducted. That’s more representative, but it also gets people around some of the other questions of authorship that would come up, such as: “Can you add somebody as a local author when you’re going into a regional meeting being held in a foreign language?” If you’ve included everybody from the beginning, you’ve already included someone from each region in your authorship team. And groups like the Asian Council of Science Editors are interested in greater involvement of voices from the global south.

LDT: What is your advice for colleagues who are interested in getting more involved in this field?

MK: Start with learning on the job. That’s first and foremost. Nothing can replace first-hand experience. Today, we have so many resources and so many tools. When a new publication professional joins an organization, they should reach out to their scientific publication teams (or to senior colleagues within their department, if it’s a smaller organization) to assimilate and learn the core competencies required to be successful in their role. Large companies have dedicated training programs for new publications professional staff, so they go through mentorship and formal assimilation programs, but that may not be the case in smaller organizations. That’s why it’s important for professionals in smaller companies to reach out to professional societies. We have DIA, we have ISMPP, we have AMWA. Resources today are much more robust and offer a lot of training and education.

LDT: One of the interesting things about the medical publication field is that it’s heavily involved in broader fields like scientific affairs or medical affairs. DIA has a medical affairs and scientific communication meeting, which is a great way to meet people in various kinds of roles. If you attend a professional society meeting, you can meet people there.

JJ: Networking is very important. People move in and out of the writing role. A lot of people I know started out in other roles but were drawn in to write because their group needed a writer, and they said, “I love this work, I’m going to stick in this role.” I also know writers who have then used it to move to jobs in medical affairs, in HEOR (health economics and outcomes research), and other things. Personally, I have been a writer my entire career. It’s what I love. It’s where I’ve been. But a lot of people pass through here on different career paths, and that’s great.

MK: If we look back at the industry publications landscape, we still have many challenges. But have we moved the needle? We have moved the needle across many different topics in the publications and medical writing fields, and we should all feel proud of what we have accomplished. There’s so much more to come. But it’s a great place to be.

LDT: I agree. People have worked to build a community of practice and community of medical publication professionals. If people are interested in roles in this field, it may be helpful to think about this as not just a job but as becoming part of a broader community. We’ll have to look forward to continued guidance from journal editors and regulators about scientific communication, social media, AI, and authorship.

Listen to the complete explanation of this authorship algorithm below.