NEJM AI Editorial Policies

Every person listed as an author must meet the International Committee of Medical Journal Editors (ICMJE) Recommendations for authorship. These include:

●    Substantial contributions to conception and design; or acquisition, analysis, or interpretation of data
●    Drafting of the article or critical revision for important intellectual content
●    Final approval of the version to be published
●    Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the article are appropriately investigated and resolved

First Authorship: Two or more authors may claim first/senior authorship. This should be indicated with a note indicating “Drs. A and B contributed equally to this manuscript” at the end of the author list.

Corresponding Author: Only one corresponding author is permitted for each manuscript during the submission and revision process. This author is solely responsible for all communication with NEJM AI and any associated vendors providing editing services towards acceptance and publication. After the manuscript has been accepted for publication, the authors may designate more than one person to be contacted by readers, but the original corresponding author will retain responsibilities as manuscript guarantor and will remain corresponding author in the NEJM AI records.

Authors will be asked to indicate who wrote the first draft of the manuscript. If it was not one of the authors, we will ask that authors name the person or persons and indicate who paid them. If any writing assistance other than copy editing was provided, we ask that authors name the person or persons and indicate who paid them. If authoring, artificial intelligence, machine learning, and/or large language model tools were used while composing your manuscript, we ask that those be disclosed.

●    Authors must disclose at submission of the manuscript whether AI-assisted technologies (such as large language models, chatbots, or image creators) were used to produce the submitted work. If so, both the cover letter and the submitted work should include a description of the technologies used and what was produced.

●    Because the authors of a manuscript are responsible for the accuracy, integrity, and originality of the work, chatbots or other AI-assisted technologies cannot be listed as authors.

●    Authors should carefully review and edit all materials produced through the use of AI, to prevent the submission of authoritative-sounding output that is incorrect, incomplete, or biased.

●    Authors should be able to assert that there is no plagiarism of text or images in materials produced by AI. Authors must ensure that all quoted material is properly attributed, including full citations.

●  Citation of AI-generated material as a primary source is not acceptable.

NEJM AI considers full and open disclosure to be a key aspect of the publication process.

Authors of every article type are required to complete an online disclosure form via Convey, which will be published online with the article. In addition, editors review all authors of review articles and commentaries to ensure that they are free from significant financial interests relevant to, or the appearance thereof, regarding the subject being reviewed or commented upon. These data are reviewed by the editors before authors are invited to submit their work.

In addition to the individual authors’ disclosures, each original article contains a clear statement of all sources of support for the published work, whether financial or in kind. Our policies in these matters adhere to the guidance established by the International Committee of Medical Journal Editors (ICMJE) (https://www.icmje.org/), the Committee on Publication Ethics (COPE) (https://publicationethics.org/), and the Council of Science Editors (CSE) (https://www.councilscienceeditors.org/); additional information can be found at their respective websites. Should concerns be raised about the disclosure in a given article, NEJM AI follows the guidance given by the COPE and ICMJE.

All articles published by NEJM AI describing original research carried out in human patients or in animals must have been approved by the appropriate governing bodies for the protection of research participants in the place or places where the research was performed. This includes a clear statement as to how informed consent was obtained and what body approved the informed consent document that was used in the trial. A statement indicating that this is the case must be included in the submission to NEJM AI. Editors further expect that research involving humans be conducted in accord with the World Medical Association’s Declaration of Helsinki (https://www.wma.net/policies-post/wma-declaration-of-helsinki-ethical-principles-for-medical-research-involving-human-subjects/) and that this be communicated to the editor handling the article during the review process.

NEJM AI adheres to the guidance in these matters given by ICMJE (https://www.icmje.org/), COPE (https://publicationethics.org/), and CSE (https://www.councilscienceeditors.org/); additional information can be found at their respective websites.

Misconduct: Misconduct in publishing takes many forms, including but not limited to plagiarism, fabrication or falsification of data, duplicate publication, misuse of confidential information, failure to disclose conflicts of interest, or the inclusion or exclusion of authors in a manner that does not reflect their contributions. NEJM AI takes all allegations of misconduct seriously, and reader or author concerns should be addressed to our editorial office at editorial.ai@nejm.org.

NEJM AI is committed to ensuring that any allegations of misconduct are handled in a timely and impartial manner. After reviewing allegations, the journal will determine if there is sufficient evidence to warrant an investigation. All investigations will be held to strict standards of confidentiality and fairness, and will adhere to the guidelines and principles outlined by COPE (http://publicationethics.org/resources) and ICMJE (https://www.icmje.org/recommendations/browse/publishing-and-editorial-issues/). A final determination will be communicated to all parties involved, including individuals and institutions. A finding of misconduct may result in the rejection of articles under consideration, the retraction of published articles, the notification of individuals’ home institutions, or prohibiting individuals from submitting to or otherwise participating in NEJM AI or other NEJM Group publications.

Appeals and Complaints: NEJM AI strives to provide a safe and transparent platform where feedback is welcome from authors, reviewers, and the broader research community. The Journal’s standard operations provide regular opportunities for feedback during and after review and production, and critical feedback is welcomed as we continue to improve our author and reviewer experience.

Individuals who wish to formally appeal Editor or Journal decisions may submit an appeal letter to the NEJM AI editorial office at editorial.ai@nejm.org. Letters should explain clearly the basis for a disagreement with a decision or outcome and should provide any information or evidence that the Journal should consider for the appeal. All appeals will be considered in a timely manner, notwithstanding the need to conduct additional review or gather further information from the appellant, and a final determination will be communicated to the appellant as soon as is feasible.

Complaints about publication ethics, processes, or outcomes should be addressed to the NEJM AI editorial office at editorial.ai@nejm.org. The procedure that follows will depend on the basis and nature of the complaint. Complaints about processes may produce feedback to improve peer review or manuscript production; complaints regarding ethics may result in formal misconduct investigations. Regardless of the selected procedure, the Journal will provide feedback and a planned course of action to the complainant in a timely manner.

●    NEJM AI conforms to publication and review standards set out by organizations dedicated to improving the reliability, transparency and utility of published information. In this regard we follow the policies outlined in the Principles of Transparency and Best Practice in Scholarly Publishing (last updated September 15, 2022) which can be found at https://publicationethics.org/resources/guidelines/principles-transparency-and-best-practice-scholarly-publishing#:~:text=The%20Principles%20of%20Transparency%20and,that%20the%20journal%20follows.ws. This joint statement by COPE, the Directory of Open Access Journals (DOAJ), the Open Access Scholarly Publishing Association (OASPA), and the World Association of Medical Editors (WAME) identifies and codifies best practice publication policies.
●    NEJM AI follows the guidelines of COPE and fully cooperates with its investigations when issues arise with respect to publication ethics.
●    We also adhere to the guidelines promulgated by ICMJE and CSE. These organizations set standards and best practices for medical and scientific publishing.
●    NEJM AI articles are published in accord with standards promulgated by National Information Standards Organization (NISO) (https://www.niso.org/) including its committee promulgating practices for e-journals known as Recommended Practices for the Presentation and Identification of E-Journals (PIE-J) (https://www.niso.org/standards-committees/pie-j).
●    NEJM AI archives its materials via Portico so that they will be available in perpetuity.

All original research articles are peer-reviewed and require a structured (i.e., Background, Methods, Results, and Conclusions) abstract. Authors must report key findings in the abstract in the form of key data and not make unsupported statements.

Articles other than original research articles must contain a non-structured abstract. This narrative abstract should identify the key points made in the article and the basis of support for these key points.

All articles must also include a short description. Content from the abstract may be used in the description.

All articles that have received funding from an entity in support of the work being presented must provide a support statement in the manuscript.

ICMJE and NEJM AI require investigators to submit a data-sharing statement (2018) and register a data-sharing plan when registering a trial (2019). NEJM AI will expect all authors who receive funding from the United States National Institutes of Health (NIH) to follow the NIH data sharing policies as noted in https://sharing.nih.gov/data-management-and-sharing-policy.

Moreover, editors of NEJM AI expect that the data sharing policies and practices for all original research articles funded from sources other than the NIH will conform to the policies of the respective funders and the countries of origin of the data.

NEJM AI is committed to fostering efforts to ensure that study participants are representative of the patients affected by the condition being studied. Clinical research studies published in NEJM AI must include in the Supplementary Appendix a table that provides background information on the disease, problem, or condition, and the representativeness of the study group. The table should be referenced within the paper, with comment (when appropriate) about the representativeness of the patient sample and the generalizability of the research findings.

More information about this policy and an example of a supplementary table providing this information are available for reference in the NEJM editorial article, Striving for Diversity in Research Studies (N Engl J Med 2021; 385:1429-1430).

If a manuscript has tables and/or figures, they should be included at the time of submission. Tables should be embedded within the manuscript and figures should be uploaded as separate files. In the event of acceptance, these tables and figures will be modified to fit NEJM AI style. We encourage figures to be original to the manuscript and not published elsewhere; for figures or tables taken from other sources it is the author’s responsibility to make sure that needed permissions are in place.

ICMJE requires that “any research project that prospectively assigns human subjects to intervention and comparison groups to study the cause-and-effect relationship between a medical intervention and a health outcome” must be registered before the start of patient enrollment.

To be acceptable, a registry must be owned by a non-for-profit entity, be publicly accessible, and contain the 20 fields required by the World Health Organization (WHO). More information about acceptable registry criteria can be found on the WHO website.

Each manuscript will be checked on submission to determine whether the study needed registration, and, if registered, whether the registration is complete and meaningful. No manuscript will enter the editorial process until it has passed this screen.

Pilot Trials: Pilot (phase 1) trials that began on or after July 1, 2008, must be registered before first patient enrollment.

Acceptable Registries: WHO maintains a list of acceptable registries. In addition, ICMJE will accept registration in any of the primary registries that participate in the WHO International Clinical Trials Registry Platform (ICTRP).

Authors should take steps to ensure that a patient cannot be identified in the manuscript or photographs being submitted. When this is not possible, the authors must provide a written release statement by the patient. Any statement and/or photo involving an identifiable patient must be reviewed and approved in writing by the individuals involved and so noted in the manuscript.

Authors should clearly report the problem statement and the clinical relevance of the study. The dataset used should be described in detail, including the source, size, and characteristics of the data, as well as the process of data collection and any preprocessing steps taken. The algorithms and models used should be elaborated, with hyperparameters and configurations specified. The evaluation metrics and validation strategies, such as cross-validation or separate test datasets, should be explicitly stated. Results should be presented comprehensively, including performance metrics, comparisons with existing methods, and any statistical analyses conducted. The authors should also discuss the limitations of their study, potential biases in the data, and the clinical implications of their findings. Ethical considerations, including patient consent and data privacy, should be addressed. Additionally, the paper should provide sufficient information for the work to be reproducible, such as code availability and detailed implementation steps. It is recommended that papers adhere to one of the relevant reporting checklists including TRIPOD-AI (https://bmjopen.bmj.com/content/11/7/e048008.abstract) or ML-CLAIM (https://www.nature.com/articles/s41591-020-1041-y).

Our Statistical Consultants recommend the following best statistical practices in manuscripts submitted to NEJM AI. We recommend that you follow them in the design and reporting of research studies.

●    The Methods section of all manuscripts should contain a brief description of sample size and power considerations for the study, as well as a brief description of the methods for primary and secondary analyses.

●    The Methods section of all manuscripts should include a description of how missing data have been handled. Unless missingness is rare, a complete case analysis is generally not acceptable as the primary analysis and should be replaced by methods that are appropriate, given the missingness mechanism. Multiple imputation or inverse probability case weights can be used when data are missing at random; model-based methods may be more appropriate when missingness may be informative. For the NEJM AI general approach to the handling of missing data in clinical trials please see Ware et al (N Engl J Med 2012;367:1353–1354).

●    Significance tests should be accompanied by confidence intervals for estimated effect sizes, measures of association, or other parameters of interest. The confidence intervals should be adjusted to match any adjustment made to significance levels in the corresponding test.

●    Unless one-sided tests are required by study design, such as in noninferiority clinical trials, all reported P values should be two-sided. In general, P values larger than 0.01 should be reported to two decimal places, and those between 0.01 and 0.001 to three decimal places; P values smaller than 0.001 should be reported as P <0.001. Notable exceptions to this policy include P values arising from tests associated with stopping rules in clinical trials or from genome-wide association studies.

●    Results should be presented with no more precision than is of scientific value and is meaningful given the available sample size. For example, measures of association, such as odds ratios, should ordinarily be reported to two significant digits. Results derived from models should be limited to the appropriate number of significant digits.

●    Original and final protocols and statistical analysis plans (SAPs) should be submitted along with the manuscript, as well as a table of amendments made to the protocol and SAP indicating the date of the change and its content.

●    The analyses of the primary outcome in manuscripts reporting results of clinical trials should match the analyses prespecified in the original protocol, except in unusual circumstances. Analyses that do not conform to the protocol should be justified in the Methods section of the manuscript. The editors may ask for additional analyses that are not specified in the protocol.

●    When comparing outcomes in two or more groups in confirmatory analyses, investigators should use the testing procedures specified in the protocol and SAP to control overall type I error — for example, Bonferroni adjustments or prespecified hierarchical procedures. P values adjusted for multiplicity should be reported when appropriate and labeled as such in the manuscript. In hierarchical testing procedures, P values should be reported only until the last comparison for which the P value was statistically significant. P values for the first nonsignificant comparison and for all comparisons thereafter should not be reported. For prespecified exploratory analyses, investigators should use methods for controlling false discovery rate described in the SAP — for example, Benjamini–Hochberg procedures.

●    When no method to adjust for multiplicity of inferences or controlling false discovery rate was specified in the protocol or SAP of a clinical trial, the report of all secondary and exploratory endpoints should be limited to point estimates of treatment effects with 95% confidence intervals. In such cases, the Methods section should note that the widths of the intervals have not been adjusted for multiplicity and that the inferences drawn may not be reproducible. No P values should be reported for these analyses.

●    Please see Wang et al (N Engl J Med 2007;357:2189–2194) on recommended methods for analyzing subgroups. When the SAP prespecifies an analysis of certain subgroups, that analysis should conform to the method described in the SAP. If the study team believes a post hoc analysis of subgroups is important, the rationale for conducting that analysis should be stated. Post hoc analyses should be clearly labeled as post hoc in the manuscript.

●    Forest plots are often used to present results from an analysis of the consistency of a treatment effect across subgroups of factors of interest. Such plots can be a useful display of estimated treatment effects across subgroups, and the editors recommend that they be included for important subgroups. If subgroups are small, however, formal inferences about the homogeneity of treatment effects may not be feasible. A list of P values for treatment by subgroup interactions is subject to the problems of multiplicity and has limited value for inference. Therefore, in most cases, no P values for interaction should be provided in the forest plots.

●    If significance tests of safety outcomes (when not primary outcomes) are reported along with the treatment-specific estimates, no adjustment for multiplicity is necessary. Because information contained in the safety endpoints may signal problems within specific organ classes, the editors believe that the type I error rates larger than 0.05 are acceptable. Editors may request that P values be reported for comparisons of the frequency of adverse events among treatment groups, regardless of whether such comparisons were prespecified in the SAP.

●    When possible, the editors prefer that absolute event counts or rates be reported before relative risks or hazard ratios. The goal is to provide the reader with both the actual event frequency and the relative frequency. Odds ratios should be avoided, as they may overestimate the relative risks in many settings and be misinterpreted.

●    Authors should provide a flow diagram in CONSORT format. The editors also encourage authors to submit all the relevant information included in the CONSORT checklist. Although all of this information may not be published with the manuscript, it should be provided in either the manuscript or a supplementary appendix at the time of submission. The CONSORT statement, checklist, and flow diagram are available on the CONSORT website.

The validity of findings from observational studies depends on several important assumptions, including those relating to sample selection, measured and unmeasured confounding, and the adequacy of methods used to control for confounding. The Methods section of observational studies should describe how these and other relevant issues were managed in the design and analysis.

●    If an observational study included a prespecified SAP with a description of hypotheses to be tested, a signed and dated version of that plan should be included with the manuscript submission. The editors encourage authors to deposit SAPs for observational studies in one of the online repositories designed for this purpose.

●    When appropriate, observational studies should use prespecified accepted methods for controlling family-wise error rate or false discovery rate when multiple tests are conducted. In manuscripts reporting observational studies without a prespecified method for error control, summary statistics should be limited to point estimates and 95% confidence intervals. In such cases, the Methods section should note that the widths of the intervals have not been adjusted for multiplicity and that the inferences drawn from the inferences may not be reproducible. No P values should be reported for these analyses.

●    If no prespecified analysis plan exists, the Methods section should provide an outline for the planned method of analysis, including:

     -  Eligibility criteria for the selection of cases and method of sampling from the data, with a diagram as appropriate.

     -  A description of the association or causal effect to be estimated and the rationale for this choice.

     -  The prespecified method of analysis to draw inference about treatment or exposure effect or association.

●    Studies reporting the effect of a treatment or exposure should show the distribution of potential confounders and other variables, stratified by exposure or intervention group. When the analysis depends on the confounders being balanced by exposure group, differences between groups should be summarized with point estimates and 95% confidence intervals when appropriate.

●    Complex models and their diagnostics can often be best described in a supplementary appendix. Authors are encouraged to conduct an analysis that quantifies potential sensitivity to bias from unmeasured confounding; absent that, authors must provide a discussion of potential biases induced by unmeasured confounders.

●    Authors are encouraged to retest findings in a similar but independent study or studies to assess the robustness of their findings.

All manuscripts that the editors find relevant and interesting to NEJM AI are subject to a highly rigorous peer-review process, as would be expected from a publication of NEJM Group.

All original research and review articles undergo outside scientific peer review by at least two peer reviewers; original research articles and some review articles may also undergo dedicated methodological review to ensure that conclusions drawn are supported by the data provided. Authors are expected to respond to the issues raised in the peer review process.

Peer reviewers are trusted to maintain confidentiality. Peer reviewers who have or have had a close working relationship with one or more authors of the requested manuscript, are encouraged to disclose such information and recuse the role of review.

Peer Review is single-blind (i.e., the reviewer knows the identity of the authors, but the authors do not know the identity of the reviewer) and is conducted under the guidance given by ICMJE, COPE, and CSE.  

Reviewers must maintain the confidentiality of the manuscript as outlined above, which may prohibit the uploading of the manuscript to software or other AI technologies where confidentiality cannot be assured. Reviewers should disclose to journals if and how AI technology is being used to facilitate their review. Reviewers should be aware that AI can generate authoritative-sounding output that can be incorrect, incomplete, or biased.

NEJM AI editors will review the paper and if they feel that peer review is warranted, they will assign expert reviewers in the field covered by the work to examine the submission and to make recommendations. The editors will take these recommendations under consideration when they determine if the manuscript is to move forward towards publication or be declined.

Peer reviewers are reminded at the time they accept the article to review that the material they are being asked to review should be kept strictly confidential (i.e., not shared with anyone). NEJM AI will keep the identity of the peer reviewers confidential. Peer reviewers may, at their discretion, choose to reveal their identity to the author.

Reviewers who agree to review a manuscript are asked to immediately disclose any personal, professional, or financial conflicts with the authors and/or the topic in the manuscript. When a conflict arises, the editors seek out alternate reviewers.

We ask for reviews to be returned within two weeks of the reviewer agreeing to undertake the project. Reviewed materials are considered privileged information and cannot be copied or shared.

Once a manuscript has been peer-reviewed, the NEJM AI editors make the decision to either reject the submission or move the manuscript forward towards publication. A decision letter is sent to the authors detailing questions raised during the peer-review process and, where applicable, recommends revisions to meet NEJM AI publication standards. The authors in turn respond with a revised manuscript accompanied by a letter detailing all changes.

When authors resubmit their revised manuscript, it is again reviewed internally by the editor and a decision is made whether further peer or methodological review is needed. If additional outside review is not needed, the manuscript will be sent for additional editing and revisions in collaboration with the editor and its authors.

Following formal acceptance for publication, a paper will then move through the process for manuscript editing, production, and publication.

Upon submission, authors will be asked to confirm that their manuscript, in part or in whole, has not been published nor is under consideration for publication elsewhere.

NEJM AI will not consider any manuscripts that have been published elsewhere, except in the following cases:

●    Research presented at scientific meetings
●    Releasing results to government agencies to meet statutory requirements or urgent public health needs
●    Articles posted on a preprint server

When needed, corrections will be made to published articles to be sure that the published record is correct. A “Correction Notice” will be published identifying where the error lies, and the changes made to ensure the published record is correct. The article will carry a notice that a correction has been made; the original, i.e., the incorrect version of the material, can be made available upon request.

In cases in which statements in an article about how the research was performed are determined to be untrue or when the data underlying an article are found to be in question, after following the investigative guidance of ICMJE, COPE, and CSE, the published article will be labeled as retracted but maintained on the NEJM AI website.

In cases where the editors have concerns about the veracity of statements made in an article about research performance or about the integrity of the data underlying a research communication, but when a firm determination cannot be made as to these issues, the editors may post an “Expression of Concern” as noted by ICMJE. “Expression of Concern” may be withdrawn when the issues are resolved or when the article in question is retracted. In cases where the issues cannot be resolved, the “Expression of Concern” may remain in place.

NEJM AI only evaluates full manuscript submissions and does not accept presubmission inquiries.

 

Six months after publication, NEJM AI will make the full text of research articles available free of charge at ai.nejm.org for non-commercial use. Certain materials may be released sooner at the discretion of NEJM AI editors. There is no charge to submit or publish in NEJM AI.

NEJM AI uses X and LinkedIn as places where readers may post discussion about published work. The editors of NEJM AI suggest that authors monitor designated social media channels and respond as appropriate through a fortnight after formal publication on the NEJM AI website.

Material published in NEJM AI (ISSN: 2836-9386) is covered by copyright. No copyright is claimed to any work of the U.S. government.

Authors are required to sign a Publishing Agreement and Authorship Statement before a manuscript is accepted for publication. Additional information is available on the NEJM Group copyright information page.

Following publication in NEJM AI, authors are granted rights to reuse published versions of their articles as follows:

●    Share with colleagues for educational use
●    Include portions, such as figures and tables, in book chapters or other educational articles written by the author
●    Include a copy, in full or in part, in the author’s thesis or dissertation
●    Include in collections of the author’s educational writing
●    Provide copies of the article to students in classes taught by the author that have no commercial ties (i.e., those sponsored by academic institutions or scientific societies)
●    Deposit for display at the author’s academic institution’s online repository six months after publication
●    Credit should read as follows: “From [Publication Title, Author(s), Title of Article, Volume No., Page No. Copyright © (notice year) Massachusetts Medical Society. Reprinted with permission.”

For permission-related inquiries including permission to copy, reproduce, or republish content from NEJM AI, contact NEJM Group Permissions at permissions@nejm.org.

NEJM AI works with authors of articles that report research supported by funding bodies with open access mandates.

NIH-funded authors: For articles arising from research funded by the U.S. National Institutes of Health (NIH), NEJM AI will automatically submit a PDF of the published versions to the NIH Manuscript Submission System to be released six months after the article is published by NEJM AI.

While we do not yet know the specifics of the NIH and other U.S. government funding agencies’ policies in response to the 2022 Office of Science and Technology Policy (OSTP) Public Access Memorandum, NEJM AI will continue to work with authors to fulfill funder mandates, as has been our practice.

Plan-S funded authors: For articles arising from research funded by a Coalition S member, upon publication, a PDF of the author accepted manuscript (the accepted version before the final editing, enhancements, and production) with a link to the published version of the article will be made available for immediate self-archiving under the terms of the Creative Commons Attribution-No Derivatives 4.0 International license (CC-BY-ND-4.0).

Other not-for-profit funded authors: For articles arising from research funded by other not-for-profit organizations that also require submission to publicly available, not-for-profit repositories, authors are permitted to submit digital published versions (as supplied by NEJM AI in PDF format) on the following conditions:

Author requests, and not-for-profit repository agrees, that public release of the PDF will not occur until six months after the article is published.

To avoid confusion that might arise from multiple versions of the same article appearing in the medical literature, author agrees to submit only the PDF version supplied by NEJM AI to any publicly available repository.

NEJM AI receives funding from a variety of sources, including individual and institutional subscriptions, reprints and permissions requests, and third-party sponsorships, supplements, and advertising. All editorial decisions are conducted independently of funding or commercial considerations.