The January 2026 ICMJE Update: A Practical Guide for Authors

The International Committee of Medical Journal Editors updated its Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals in January 2026. If your last careful reading of the ICMJE Recommendations was sometime before that, you are now operating with an outdated map. The January update is not cosmetic. It adds two entirely new sections, revises several existing ones, and formalizes expectations that were previously left to journal-by-journal interpretation. The practical result is that certain actions now carry explicit ICMJE weight behind them in ways they did not before.

The ICMJE is not a regulatory body. It cannot compel any journal to do anything. What it does is set the normative standard that the majority of serious biomedical journals either adopt directly or treat as a baseline. If you are submitting to any journal that says its policies align with ICMJE Recommendations, and most reputable medical journals say exactly that, then the January 2026 changes are your submission requirements now. This guide walks through each major change and what it means in practice.

Why ICMJE Updates Matter More Than Most Author Guidance

Most publisher guidance is one journal's policy. ICMJE recommendations are a shared baseline that hundreds of member-aligned journals, including JAMA, the New England Journal of Medicine, The Lancet, BMJ, and their affiliated specialty publications, treat as authoritative. When ICMJE says something formally, journals that previously handled it by custom or internal policy tend to codify it. That is why an ICMJE update, even when it confirms practice that already existed informally at many journals, has downstream effects that are broader than any single publisher policy change.

The January 2026 update is also notable for its tone. Earlier iterations of the Recommendations tended to articulate principles and leave implementation details open. The 2026 revision moves toward what observers from Enago and other medical communications consultancies have called more operational accountability: specifying not just what authors should do, but in certain cases how and when they should demonstrate they have done it. That shift has direct effects on how you prepare manuscripts, how you structure collaborative agreements, and how you handle AI tools from the first draft onward.

The New AI Section: What Authors Must Now Do

Section V of the revised Recommendations is the first time ICMJE has created a standalone section devoted to artificial intelligence. It establishes two non-negotiable positions. First, AI tools and AI-assisted technologies cannot be listed as authors. This was not a surprise, given that every major publisher had already taken this position, but the ICMJE formalization removes any remaining ambiguity for journals that had not issued their own explicit policy. Second, the section mandates disclosure of AI use at the point of submission, not only in the cover letter but in the submitted work itself.

The new Section V specifies where disclosure belongs depending on the type of use. If AI was used for writing assistance, language polishing, translation, or related manuscript preparation tasks, the disclosure should appear in the acknowledgments section. If AI was used for data collection, analysis, classification, or figure generation, the disclosure should appear in the methods section, because it then describes a methodological decision that affects how readers evaluate the results. This distinction matters. Many authors who conscientiously add an AI disclosure to the acknowledgments are still non-compliant if they used a generative tool to analyze free-text responses or code qualitative data, because that use needs to be in the methods.

The section also addresses a consequence that many authors have not fully absorbed: nondisclosure of AI use may constitute misconduct in some circumstances. That phrasing is deliberate. It does not say that using AI is misconduct. It says that failing to disclose it when required is misconduct. The practical implication is that an undisclosed AI use which later comes to light, whether through a hallucinated citation, a post-publication reader complaint, or an editorial review, now carries the formal weight of a potential integrity violation rather than just a procedural error. That changes the calculus for authors who have been treating disclosure as optional or low-stakes.

There is one nuance worth noting for teams that use AI at multiple stages. The ICMJE requirement applies to AI used in the production of the submitted work. If every co-author used a grammar tool on their own drafts and the manuscript was then assembled and re-edited by a human without any AI involvement in the final version, reasonable people may disagree about what exactly must be disclosed. When in doubt, disclose at the acknowledgments level. The section explicitly frames nondisclosure as the compliance risk, not disclosure.

The New Data Access Requirement

Section III.L.2 is the change that many clinical trialists and industry-funded researchers should be paying closest attention to. It establishes a formal requirement that authors have sufficient access to the study data to take responsibility for the integrity and accuracy of their work. For single-institution academic studies, this is not a new burden. Authors at universities typically control their own datasets. The new requirement bites most clearly in two common scenarios: collaborative research that mixes academic and commercial or government contributors, and industry-sponsored trials where contractual data access for academic investigators is limited by design.

The section specifies that in collaborative research involving both academic and non-academic contributors, at least one author affiliated with an academic institution should have direct access to the primary dataset and play an active role in the analysis. This sentence is worth reading carefully. It does not say an academic author reviewed summaries, was briefed on results, or approved the final tables. It says direct access to the primary dataset and active analytical involvement. That distinction rules out the compliance model in which an academically credentialed author is listed to provide institutional legitimacy while the underlying data remains inside a sponsor's database.

This requirement did not appear from nowhere. It responds to a well-documented pattern in industry-sponsored clinical research where published conclusions have later been found inconsistent with the underlying trial data, sometimes because academic authors had access only to summary statistics or sponsor-prepared tables rather than individual-level data. High-profile examples from cardiology and psychiatry over the past decade established that limited author access to primary data was not merely an inconvenience but a precondition for certain kinds of result manipulation. The ICMJE has been building toward this requirement for some years; Section III.L.2 makes it explicit.

Updated Sponsor Agreement Guidance

Section II.B, which covers sponsor agreements and academic freedom, was revised alongside the new data access section. The 2026 language reinforces that contractual agreements between researchers and commercial sponsors must not restrict the authors' ability to access data, conduct independent analysis, or publish findings without sponsor approval. This is not a new principle in academic medicine, but the revision tightens the connection between these contractual expectations and the requirements that journals can impose at submission.

For authors preparing manuscripts from externally funded work, the practical implication is that journal editors aligned with ICMJE now have formal backing to ask whether your research agreement meets this standard. Some journals have already begun to include a declaration in their submission system asking authors to confirm that the sponsor had no role in data analysis or publication decisions, or to describe the sponsor's role explicitly if it did. That kind of declaration is now grounded in ICMJE guidance rather than individual journal policy, which means more journals will adopt it.

Authors who have legitimate, transparent commercial funding arrangements, where the sponsor provided funding and access without controlling interpretation or publication, have nothing to fear from these revisions. The point is transparency and actual independence, not the appearance of it.

Clinical Trial Registration: Stricter Verification

Section III.L.1 on clinical trial registration was revised to sharpen both the timing requirement and the verification expectation. The Recommendations now advise editors to require that clinical trials be registered in publicly accessible registries meeting WHO or ICMJE standards at or before the time of first participant consent for enrollment. The prior standard was registration before the start of participant enrollment, which left some ambiguity about pre-enrollment preparation phases. The new language anchors the timing to first participant consent.

More consequentially, the revision now encourages editors to actively verify that registration is complete, accurate, and conducted before participant enrollment began. This is a shift from trusting the registration number that appears in a manuscript to checking the registration record itself. For authors, this means that incomplete trial registrations, registrations filed after recruitment began, or registration records that do not match the methods as submitted are now formal compliance issues at the editorial level, not just background concerns about transparency.

This last point, consistency between registration and manuscript, matters most for prospective trialists who adjust their primary outcome after seeing interim results and before publishing. ICMJE has always opposed outcome switching, but the new verification emphasis gives editors more formal standing to flag discrepancies they find during manuscript handling. If your registered and reported outcomes differ, you should address this explicitly in your manuscript, typically in the methods section, rather than hoping it is not noticed.

Revised Scientific Misconduct Definitions and Correction Terminology

Section III.B on scientific misconduct received terminological refinements in the 2026 update. The revisions clarify the hierarchy of post-publication corrections available to journals: corrections for minor errors, expressions of concern when integrity questions cannot be resolved quickly, and retractions for work that is sufficiently flawed that its findings cannot be relied upon. These categories are not new, but the 2026 language gives more precise guidance on when each category is appropriate and what obligation journals have to act in each case.

For authors, the most practical aspect of this revision is the clarification around expressions of concern. Some journals have historically used expressions of concern as a holding mechanism while investigations were ongoing, then quietly removed them without further notice. The updated guidance implies a cleaner resolution requirement: an expression of concern should eventually conclude with either a retraction, a correction, or a formal editorial statement that the investigation found no problem. Authors who have open expressions of concern sitting against their published work may want to contact the editorial office of the relevant journal to understand where that process stands, since ICMJE-aligned journals now have clearer obligations to resolve them.

The misconduct section also now links explicitly to the new AI disclosure provision, making clear that nondisclosure of AI use is a form of research misconduct rather than simply a breach of editorial policy. The distinction matters for institutional consequences. A finding of policy breach stays at the journal level. A finding of misconduct can be referred to an author's institution and becomes part of the public record.

Sources of Support: Updated Transparency Requirements

Section IV.3.A on reporting sources of support was updated to align with the broader transparency themes of the 2026 revision. Authors are now expected to report all sources of financial support for the research, including indirect support such as salary contributions from institutional grants and equipment or software provided by commercial entities, not only direct research funding. This is broader than what many researchers are used to disclosing. If a pharmaceutical company provided drug supply at no cost, or a medical device company provided equipment, those contributions now qualify as sources of support that belong in the funding statement.

The updated guidance also asks authors to describe the specific role each funder played, rather than simply listing names and grant numbers. A funding statement that says "This work was funded by Grant XYZ from Agency ABC" is now less complete than one that adds "The funder had no role in study design, data collection, analysis, interpretation, or the decision to submit." Most authors at research universities are already accustomed to writing that second sentence, but for teams publishing from clinical practice settings or without research administration support, this level of specificity may be less automatic.

How Most ICMJE-Aligned Journals Will Implement This

ICMJE sets the standard; individual journals implement it at their own pace. In practice, large publishers like Elsevier, Springer Nature, Wiley, and BMJ typically update their submission system checklists and instructions for authors within a few months of an ICMJE revision. Smaller independent journals can take longer, and specialty society journals sometimes take a full publishing year to incorporate changes at the submission level. This means that right now, in mid-2026, the January update is likely visible in the submission systems of major journals but may not yet have reached every specialty outlet.

The safest approach is to apply the ICMJE standard regardless of whether the specific journal has updated its checklist. If your submission goes to a journal that follows ICMJE Recommendations, and the editor flags something during peer review or revision that traces back to the January 2026 update, saying that the instructions for authors had not been revised yet is not a credible defense. The ICMJE Recommendations are publicly available and updated in January. Authors who submit in May or June of the same year are expected to have read them.

If you are genuinely uncertain whether a specific journal follows ICMJE, the most reliable signal is a statement in the instructions for authors that the journal's policies conform to or are based on ICMJE Recommendations. Almost all journals indexed in MEDLINE and many indexed in Scopus make such a statement. Journals that do not are either specialty publications with their own independent standards or journals that have not aligned with ICMJE at all, which itself is worth noting when you are evaluating publication venues.

A Pre-Submission Checklist for the 2026 Requirements

Before submitting any manuscript to an ICMJE-aligned medical journal in 2026, it helps to walk through the January update systematically. None of the items below are bureaucratic formalities. Each one either connects to a new formal requirement or to an existing requirement that the 2026 revision has made more explicit.

Journals are increasingly building these checks into their submission systems directly. You may encounter a structured declaration form that asks about each of these areas, particularly for the AI and trial registration points. Having clean answers before you reach that form is faster and creates fewer revision cycles.

What This Means for Researchers at Earlier Career Stages

Graduate students, residents, fellows, and early-career faculty often inherit authorship practices from supervisors and research groups whose habits formed when many of these requirements were less formal. The January 2026 revision is a reasonable moment to audit those inherited habits. If your research group uses AI tools as a matter of routine but has never established a disclosure protocol, this update is the practical reason to create one now, before a submission goes out without it.

The data access requirement is particularly worth raising with supervisors in clinical research groups that collaborate with industry partners. Postdoctoral researchers and graduate students almost never negotiate research agreements themselves, but they can and should ask their PI whether the agreement governing a collaborative study meets the ICMJE data access standard. If it does not, that is a question of both scientific integrity and personal career exposure, since the corresponding author's name is not the only one that appears on a retraction notice.

The clinical trial registration revision is most relevant to trainees running prospective studies within larger research programs. Registration before first participant consent is not always the first thing on a trainee's list during a busy study startup. Making it that is now the correct approach, and the January 2026 ICMJE language gives residents and fellows a defensible basis for pushing back if a supervisor suggests it can be done later.

Reading the ICMJE Recommendations Directly

The ICMJE publishes its Recommendations in full at icmje.org, where the document is freely accessible both as a PDF and in a browseable HTML format organized by section. The January 2026 update is accompanied by a brief editorial on the ICMJE news and editorials page that summarizes what changed and explains the rationale for the new additions. Reading the editorial first, then navigating to the specific new and revised sections, is faster than reading the entire document from scratch.

The Recommendations are updated periodically, and ICMJE does not send notifications to individual authors or institutions. The practical habit is to check the news page before any major submission, particularly if six months or more have passed since your last review of the document. For most research groups, building a quick annual ICMJE review into the beginning of the academic year is a low-overhead way to stay current. The January timing of most ICMJE updates makes early spring a natural moment for this.

No document, including this one, is a substitute for reading the source. Summaries compress and occasionally simplify. Editors and reviewers work from the primary text. If a specific point is in dispute during peer review or revision, the authoritative answer comes from the Recommendations themselves, not from a third-party interpretation of them.

Further Reading