CRediT Author Contribution Statements: A Practical Guide for Medical Researchers in 2026

Until recently, the author contribution statement at the end of a medical paper was easy to ignore. Teams would copy a previous paper's contributions block, shuffle the names to reflect rough seniority, and paste it into the submission system without much thought. Reviewers did not check it. Editors rarely questioned it. The statement existed largely to satisfy a formal requirement, not to communicate anything specific.

That posture is becoming harder to maintain. Elsevier, Springer Nature, Wiley, PLOS, eLife, Taylor and Francis, SAGE, and BMC all require CRediT-formatted contribution statements through their submission portals as of 2025, and most have made it a mandatory field rather than an optional one. In 2026, Crossref began embedding CRediT metadata in its public database, which means these statements are no longer buried on the last page of a PDF. They are now indexed, searchable, and attached to your DOI. What you write in that box is part of the permanent record of what you claim to have done.

At the same time, the National Institutes of Health Intramural Research Program published a 2025 paper describing how it uses CRediT as a structured tool to mediate authorship disputes before and after publication. The taxonomy has become infrastructure, not just etiquette. Medical researchers who fill it in carelessly are taking a real risk, and those who fill it in thoughtfully are creating documentation that protects them if authorship becomes contested later.

Why Authorship Needed a More Precise Language

The International Committee of Medical Journal Editors has defined authorship through four criteria since the 1980s. To be listed as an author, a person must have made substantial contributions to the conception or design of the work, or to the acquisition, analysis, or interpretation of data. They must have participated in drafting or critically revising the manuscript for important intellectual content. They must have given final approval of the version being submitted. And they must agree to be accountable for all aspects of the work, including the integrity of any part they did not directly handle.

These four ICMJE criteria are principled and well-reasoned. They are also written at a level of generality that makes them easy to satisfy on paper while violating their spirit in practice. A lab director who reviews a draft once and gives final approval meets all four criteria. A postdoc who collected the entire dataset, ran all the analyses, and wrote the first three drafts of the manuscript also meets all four criteria. The authorship line treats them identically. It usually does not tell a reader, a hiring committee, or an integrity investigator who actually did the work.

That gap was the motivation for CRediT. The taxonomy grew from a 2012 workshop convened jointly by Harvard University and the Wellcome Trust, with input from researchers, publishers, and ICMJE representatives. The first version was published in 2014. In 2022 it became an ANSI/NISO standard (Z39.104-2022). What it provides is a controlled vocabulary of exactly what people actually do on papers, broken into 14 defined roles that are specific enough to be useful and general enough to apply across fields.

The 14 CRediT Roles, Explained for Clinical Research

The taxonomy was designed to be self-explanatory, and most of the roles read as obvious once you see them. Where medical research teams run into trouble is usually at the edges: roles that sound similar to each other, roles whose scope they underestimate, and roles that apply to people they had not considered listing at all.

Each author should be assigned one or more roles, with a designation of whether their contribution was lead, supporting, or equal. The statement usually appears at the end of the manuscript or in a dedicated submission system field. In practice, most publishers display it in a standard format: Author A: Conceptualization, Investigation, Writing — original draft. Author B: Formal analysis, Visualization, Writing — review and editing.

One thing the taxonomy does not do is replace ICMJE authorship criteria. CRediT and ICMJE work together. The ICMJE criteria answer the threshold question of who qualifies to be listed as an author at all. CRediT answers the finer question of what each author actually did. A person who contributed only resources, or only funding acquisition, or only project administration, is a valuable contributor but may not satisfy all four ICMJE criteria for authorship. Those contributors belong in the acknowledgements section, with their specific contributions noted separately.

Why the Crossref 2026 Integration Changes the Stakes

For most of its existence, CRediT lived inside PDFs. A reader could find the contribution statement if they scrolled to the end of the paper, but the information was not machine-readable in any systematic way. That has begun to change. Crossref, which assigns and maintains digital object identifiers for the majority of scholarly publications, started incorporating CRediT metadata into its database schema in 2026. This means that when a publisher deposits metadata for a new article, they can now also deposit the structured CRediT roles for each named author, and those roles become queryable through Crossref's API.

The practical consequence is not immediately dramatic. Most researchers are not running API queries against Crossref to check what roles people claimed. But the infrastructure matters over time. Funders can ask how many papers a researcher has led as writing — original draft author versus how many they have contributed to only as a reviewer. Institutions can examine what role a supervisor actually played across a team's output. Journals can flag discrepancies when the same person appears as lead author on the conceptualization for fifty papers in a single year. The Crossref integration turns CRediT from a self-reported narrative into indexed, cross-comparable data.

For individual researchers, the implication is straightforward: what you assert in your CRediT statement is now part of your searchable scholarly identity, not just a box you filled in for submission compliance. This is reason enough to take the statement seriously before you submit rather than approving whatever a co-author drafts without checking.

Which Publishers Now Require CRediT, and Where the Mandate Sits

The adoption landscape has consolidated considerably since 2020. Elsevier began requiring CRediT author statements across its journals in 2019, integrated into the Editorial Manager submission system. Springer Nature and BMC journals have required it for several years, with the field built into their submission portals. Wiley, Taylor and Francis, SAGE, and PLOS have all moved to mandatory CRediT across the majority of their journal portfolios. eLife has required it since 2018. At Oxford University Press and Cambridge University Press, adoption varies by journal, but most biomedical titles in both portfolios now include the field.

The flagship general medical journals have their own approaches. The New England Journal of Medicine, JAMA, BMJ, and The Lancet all require author contribution statements, though their specific formatting differs from standard CRediT in some cases. The BMJ uses a long-standing author contribution format that predates CRediT and does not map directly to the taxonomy's 14 roles; it asks for free-text descriptions of each author's role. JAMA and The Lancet use their own templates that overlap substantially with CRediT categories but may not use the exact taxonomy headings. If you are targeting one of these journals, read the current instructions for authors rather than assuming a CRediT statement will transfer without revision.

For journals using Editorial Manager (the most widely deployed manuscript management platform in biomedical publishing), the CRediT roles typically appear as a step in the submission workflow where each listed author's contributions are specified through a checkbox interface. This is the moment where many teams run into problems, because the person doing the submission fills in the checkboxes from memory or assumption rather than from a pre-agreed record of what each person actually did.

How NIH Is Using CRediT to Resolve Authorship Disputes

Authorship disputes are common enough in research institutions that many have established formal processes for handling them, but the processes often lacked a consistent framework for describing contributions. The NIH Intramural Research Program's 2025 paper describing its use of CRediT as a mediation tool is worth understanding because the NIH IRP handles disputes across hundreds of labs in clinical and biomedical research, which makes their experience broadly applicable.

The core approach is to ask each party in a dispute to independently assign CRediT roles to every person involved in the project, then compare the resulting maps. When two team members disagree about authorship, their assessments of who did what often diverge in specific, documentable ways. Person A believes they led the conceptualization. Person B believed they led it. The taxonomy gives mediators a shared vocabulary to surface the disagreement precisely rather than arguing at the level of “I contributed more than you.”

The method also helps resolve disputes that arise post-publication, where a former team member claims they should have been listed as an author. Having a contemporaneous record of who was assigned which roles at the time of submission provides evidence that is far more reliable than reconstructed recollections from multiple parties months or years later. Several research universities in the United States have begun adopting similar CRediT-based documentation as part of their authorship agreement forms, signed at the start of a project rather than at submission.

Ghost Authorship, Gift Authorship, and the New Accountability

Two authorship problems have plagued biomedical publishing for decades. Gift authorship is when someone is listed as an author who did not meaningfully contribute to the work, usually a department head, a senior clinician, or an industry sponsor whose name adds prestige or secures funding relationships. Ghost authorship is when someone who did substantial work on the paper, often a medical writer contracted by a pharmaceutical company, is deliberately omitted from the author list to obscure the source of the manuscript.

CRediT does not make these practices impossible, but it raises the cost of both. A department head listed on thirty papers in a year who claims lead contributions in Conceptualization and Writing — original draft across all thirty creates a record that is increasingly implausible to defend, particularly once that data sits in Crossref and can be compared across papers. A paper whose writing — original draft role is assigned to no one, or whose acknowledgements section lists a medical writer without a corresponding CRediT entry, creates a gap that editors and post-publication reviewers are increasingly trained to notice.

The ICMJE has addressed ghost authorship directly in its recommendations by specifying that anyone who contributed substantially to the work but is not listed as an author must appear in the acknowledgements section, with a description of their specific contribution. Medical writers who draft or substantially edit a manuscript should be acknowledged by name and their role described. CRediT-style role language in the acknowledgements section, even for non-authors, is increasingly common as a way to satisfy this requirement clearly.

How to Fill In Your CRediT Statement Correctly

The most common error is a single person, usually the corresponding author or the most junior member of the team, completing the CRediT form on behalf of everyone without consulting the other authors. This creates two problems: the roles may be inaccurate, and the authors who did not review their own entries have not actually approved what their name is now attached to. Treat the CRediT statement like you treat authorship itself. Every listed author should see their own CRediT entry before the paper is submitted.

The process works best when the corresponding author sends a draft of the full contribution table to all co-authors and asks each person to review their own row and raise any concerns. This does not need to be a formal meeting. A reply-all email with the table attached, and a request for corrections within a specified deadline, is sufficient. Keep the email thread. If a dispute arises later, that thread is evidence that the statement was reviewed and accepted by each author before submission.

For the roles themselves, be specific rather than generous. Do not assign Conceptualization to someone who endorsed the study idea at a department meeting but did not shape the research question. Do not assign Formal analysis to someone who reviewed the analysis after the fact but did not design or conduct it. The distinction between Writing — original draft and Writing — review and editing matters and should reflect what actually happened. A statistician who wrote the methods subsection describing their approach contributed to the original draft; a co-author who added comments to a PDF revision contributed to review and editing.

Authors can hold multiple roles, and most do on a standard clinical study. A first-author PhD student on a prospective cohort paper might legitimately claim: Conceptualization (lead), Data curation (lead), Formal analysis (lead), Investigation (supporting), Methodology (supporting), Validation (supporting), Visualization (lead), Writing — original draft (lead), Writing — review and editing (supporting). A senior clinician whose patients were the study population and who reviewed drafts might claim: Resources (lead), Supervision (supporting), Writing — review and editing (supporting). Both are accurate. The taxonomy accommodates them without anyone needing to appear more prominent than their actual contribution warrants.

Gray Areas Worth Thinking Through Before You Submit

Several categories of contributors create genuine uncertainty in clinical research teams. Patient partners and public representatives who contributed to study design or outcome selection present a challenge that the taxonomy does not fully resolve. Under ICMJE criteria, they may not qualify as authors because they are not typically in a position to approve the final submitted manuscript or to be accountable for its integrity in the way clinical researchers are. But listing them only in the acknowledgements without describing what they actually contributed to is equally unsatisfying, particularly for journals that publish patient-centered research. A reasonable approach is to describe their specific contributions in the acknowledgements using CRediT role language, without formally listing them in the author block.

Biostatisticians who are contracted for a single paper rather than embedded in the research team occupy a similar gray zone. If the statistician designed the analytical approach, contributed to the methods section, and verified the analysis, that is Methodology, Formal analysis, Validation, and Writing — review and editing: four roles that amount to substantial intellectual contribution. Under ICMJE criteria, they may well qualify as authors and should be offered authorship, not just acknowledged. Many statistical consulting services actively discourage their staff from accepting authorship as a matter of policy, but that does not change the ethical analysis of what the contribution warranted.

AI tools deserve particular attention in this context. No current AI system can satisfy all four ICMJE authorship criteria. AI tools cannot take accountability for the integrity of a paper, cannot approve a final version in any meaningful sense, and are not persons who can be held responsible for errors. They therefore cannot be listed as authors, and no CRediT role should be assigned to them. But if an AI tool contributed to Data curation, Formal analysis, or Writing — original draft in a way that would require disclosure elsewhere in the manuscript, that use still belongs in the AI disclosure statement rather than in the CRediT author contribution block. The two statements address different questions.

What the Statement Looks Like in a Published Paper

Most journals publish the CRediT statement in a section titled Author contributions or Authorship contributions, typically appearing after the conclusion and before the acknowledgements. The format is usually a simple list of initials or names followed by their assigned roles. Some journals display only the role labels; others display both the label and a brief description. A few journals, particularly those in the open science tradition, display the full CRediT schema with each of the 14 roles listed and the contributing authors named for each one, which is more informative but takes up more space.

The published version matters because it is what readers, reviewers, hiring committees, and post-publication reviewers will see. If the statement contradicts what the paper describes, that discrepancy is now visible. A paper that describes a complex multi-step statistical analysis in the methods section but assigns Formal analysis to a single senior author who is also listed as supervising six other papers published in the same year draws attention. Not necessarily to misconduct, but to a story that may need explaining.

A Practical Approach for Your Next Submission

Start the conversation about contributions before the paper exists. When a project begins, sketch out who will do what using the CRediT roles as a checklist. This is not a binding contract, because research never goes exactly to plan, but it gives the team a shared vocabulary from the beginning. When the project is complete and the manuscript is drafted, revisit the original sketch and update it to reflect what actually happened. If someone's role changed, the final statement should reflect the final state.

At the point of submission, assign the CRediT roles with the same care you would give any other method section. Check the target journal's specific requirements, because the BMJ, JAMA, NEJM, and Lancet each have their own formatting conventions that differ from the standard CRediT template. Send the completed contribution table to all co-authors for explicit review before submitting. Keep a record of that review.

These steps take longer than copying a previous paper's statement and adjusting the names. They are worth the time. The landscape of academic publishing is moving toward greater individual-level accountability in authorship. The Crossref metadata integration is part of that movement. The NIH's use of CRediT in dispute mediation is part of it. The funder scrutiny of who actually produced the output they funded is part of it. A carelessly written contribution statement is not going to cause a problem on most papers, most of the time. On the paper where it does cause a problem, having the right documentation in place is the difference between a difficult conversation and a retraction inquiry.

Further Reading