Writing and Editing for Biomedical Publication

Uniform Requirements for Manuscripts Submitted to BiomedicalJournals: Writing and Editing for Biomedical PublicationUpdated April 2010

Publication Ethics: Sponsorship, Authorship, and Accountability

International Committee of Medical Journal Editors

The following information is available to be viewed/printed in Adobe Acrobat pdf format.

I. Statement of PurposeA. About the Uniform RequirementsB. Potential Users of the Uniform RequirementsC. How to Use the Uniform Requirements

II. Ethical Considerations in the Conduct andReporting of ResearchA. Authorship and Contributorship

1. Byline Authors2. Contributors Listed in Acknowledgments

B. Editorship1. The Role of the Editor2. Editorial Freedom

C. Peer ReviewD. Conflicts of Interest

1. Potential Conflicts of Interest Related toIndividual Authors’ Commitments

2. Potential Conflicts of Interest Related toProject Support

3. Potential Conflicts of Interest Related toCommitments of Editors, Journal Staff, orReviewers

E. Privacy and Confidentiality1. Patients and Study Participants2. Authors and Reviewers

F. Protection of Human Subjects and Animals inResearch

III. Publishing and Editorial Issues Related toPublication in Biomedical JournalsA. Obligation to Publish Negative StudiesB. Corrections, Retractions, and “Expressions of

Concern”C. CopyrightD. Overlapping Publications

1. Duplicate Submission2. Redundant Publication3. Acceptable Secondary Publication4. Competing Manuscripts Based on the Same

Studya. Differences in Analysis or Interpretationb. Differences in Reported Methods or

Results5. Competing Manuscripts Based on the Same

Database

E. CorrespondenceF. Supplements, Theme Issues, and Special SeriesG. Electronic PublishingH. AdvertisingI. Medical Journals and the General MediaJ. Obligation to Register Clinical Trials

IV. Manuscript Preparation and SubmissionA. Preparing a Manuscript for Submission to

Biomedical Journals1. a. General Principles

b. Reporting Guidelines for Specific StudyDesigns

2. Title Page3. Conflict-of-Interest Notification Page4. Abstract and Key Words5. Introduction6. Methods

a. Selection and Description ofParticipants

b. Technical Informationc. Statistics

7. Results8. Discussion9. References

a. General Considerations Related toReferences

b. Reference Style and Format10. Tables11. Illustrations (Figures)12. Legends for Illustrations (Figures)13. Units of Measurement14. Abbreviations and Symbols

B. Sending the Manuscript to the Journal

V. ReferencesA. Print References Cited in this DocumentB. Other Sources of Information Related to

Biomedical Journals

VI. About the International Committee of MedicalJournal Editors

VII. Authors of the Uniform Requirements

VIII. Use, Distribution, and Translation of the UniformRequirements

IX. Inquiries

1

I. STATEMENT OF PURPOSE

I. A. About the Uniform RequirementsA small group of editors of general medical journals

met informally in Vancouver, British Columbia, in 1978to establish guidelines for the format of manuscripts sub-mitted to their journals. This group became known as theVancouver Group. Its requirements for manuscripts, in-cluding formats for bibliographic references developed bythe National Library of Medicine (NLM), were first pub-lished in 1979. The Vancouver Group expanded andevolved into the International Committee of Medical Jour-nal Editors (ICMJE), which meets annually. The ICMJEhas gradually broadened its concerns to include ethicalprinciples related to publication in biomedical journals.

The ICJME has produced multiple editions of theUniform Requirements for Manuscripts Submitted to Bio-medical Journals. Over the years, issues have arisen that gobeyond manuscript preparation, resulting in developmentof a number of Separate Statements on editorial policy.The entire Uniform Requirements document was revisedin 1997; sections were updated in May 1999 and May2000. In May 2001, the ICMJE revised the sections relatedto potential conflict of interest. In 2003, the committeerevised and reorganized the entire document and incorpo-rated the Separate Statements into the text. The committeeprepared this revision in 2010.

The total content of the Uniform Requirements forManuscripts Submitted to Biomedical Journals may be re-produced for educational, not-for-profit purposes withoutregard for copyright; the committee encourages distribu-tion of the material.

Journals that agree to use the Uniform Requirementsare encouraged to state in their Instructions to Authorsthat their requirements are in accordance with the Uni-form Requirements and to cite this version. Journals thatwish to be listed on www.ICMJE.org as a publication thatfollows the Uniform Requirements should contact theICMJE secretariat office.

The ICMJE is a small working group of general med-ical journals, not an open-membership organization. Occa-sionally, the ICMJE will invite a new member or guestwhen the committee feels that the journal or organizationwill provide a new perspective. Open membership organi-zations for editors and others in biomedical publicationinclude the World Association of Medical Editors www.WAME.org, the Council of Science Editors (www.councilscienceeditors.org/), and the European Associationof Science Editors (www.ease.org.uk).

I. B. Potential Users of the Uniform RequirementsThe ICMJE created the Uniform Requirements pri-

marily to help authors and editors in their mutual task ofcreating and distributing accurate, clear, easily accessiblereports of biomedical studies. The initial sections addressthe ethical principles related to the process of evaluating,improving, and publishing manuscripts in biomedical jour-

nals and the relationships among editors and authors, peerreviewers, and the media. The latter sections address themore technical aspects of preparing and submitting manu-scripts. The ICMJE believes that the entire document isrelevant to the concerns of both authors and editors.

The Uniform Requirements can provide many otherstakeholders—peer reviewers, publishers, the media, pa-tients and their families, and general readers—with usefulinsights into the biomedical authoring and editing process.

I. C. How to Use the Uniform RequirementsThe Uniform Requirements state the ethical principles

in the conduct and reporting of research and provide rec-ommendations relating to specific elements of editing andwriting. These recommendations are based largely on theshared experience of a moderate number of editors andauthors, collected over many years, rather than on the re-sults of methodical, planned investigation that aspires to be“evidence-based.” Wherever possible, recommendationsare accompanied by a rationale that justifies them; as such,the document serves an educational purpose.

Authors will find it helpful to follow the recommen-dations in this document whenever possible because, asdescribed in the explanations, doing so improves the qual-ity and clarity of reporting in manuscripts submitted toany journal, as well as the ease of editing. At the same time,every journal has editorial requirements uniquely suited toits purposes. Authors therefore need to become familiarwith the Instructions to Authors specific to the journalthey have chosen for their manuscript—for example, thetopics suitable for that journal, and the types of papers thatmay be submitted (for example, original articles, reviews,or case reports)—and should follow those instructions.

II. ETHICAL CONSIDERATIONS IN THE CONDUCT AND

REPORTING OF RESEARCH

II A. Authorship and ContributorshipII. A. 1. Byline Authors

An “author” is generally considered to be someonewho has made substantive intellectual contributions to apublished study, and biomedical authorship continues tohave important academic, social, and financial implications(1). An author must take responsibility for at least one com-ponent of the work, should be able to identify who is respon-sible for each other component, and should ideally be confidentin their co-authors’ ability and integrity. In the past, readerswere rarely provided with information about contributionsto studies from persons listed as authors and in Acknowl-edgments (2). Some journals now request and publish in-formation about the contributions of each person named ashaving participated in a submitted study, at least for orig-inal research. Editors are strongly encouraged to developand implement a contributorship policy, as well as a policyon identifying who is responsible for the integrity of thework as a whole.

While contributorship and guarantorship policies ob-

Uniform Requirements for Manuscripts Submitted to Biomedical Journals

2 www.icmje.org

viously remove much of the ambiguity surrounding contri-butions, they leave unresolved the question of the quantityand quality of contribution that qualify for authorship.The ICJME has recommended the following criteria forauthorship; these criteria are still appropriate for journalsthat distinguish authors from other contributors.

● Authorship credit should be based on 1) substantialcontributions to conception and design, acquisition ofdata, or analysis and interpretation of data; 2) drafting thearticle or revising it critically for important intellectualcontent; and 3) final approval of the version to be pub-lished. Authors should meet conditions 1, 2, and 3.

● When a large, multicenter group has conductedthe work, the group should identify the individuals whoaccept direct responsibility for the manuscript (3). Theseindividuals should fully meet the criteria for authorship/contributorship defined above, and editors will ask theseindividuals to complete journal-specific author andconflict-of-interest disclosure forms. When submitting amanuscript authored by a group, the corresponding authorshould clearly indicate the preferred citation and identifyall individual authors as well as the group name. Journalsgenerally list other members of the group in the Acknowl-edgments. The NLM indexes the group name and thenames of individuals the group has identified as being di-rectly responsible for the manuscript; it also lists the namesof collaborators if they are listed in Acknowledgments.

● Acquisition of funding, collection of data, or gen-eral supervision of the research group alone does not con-stitute authorship.

● All persons designated as authors should qualify forauthorship, and all those who qualify should be listed.

● Each author should have participated sufficiently inthe work to take public responsibility for appropriate por-tions of the content.

Some journals now also request that one or more au-thors, referred to as “guarantors,” be identified as the per-sons who take responsibility for the integrity of the work asa whole, from inception to published article, and publishthat information.

Increasingly, authorship of multicenter trials is attrib-uted to a group. All members of the group who are namedas authors should fully meet the above criteria for author-ship/contributorship.

The group should jointly make decisions aboutcontributors/authors before submitting the manuscript forpublication. The corresponding author/guarantor shouldbe prepared to explain the presence and order of theseindividuals. It is not the role of editors to make authorship/contributorship decisions or to arbitrate conflicts related toauthorship.

II. A. 2. Contributors Listed in Acknowledgments

All contributors who do not meet the criteria for au-thorship should be listed in an acknowledgments section.Examples of those who might be acknowledged include a

person who provided purely technical help, writing assis-tance, or a department chairperson who provided only gen-eral support. Editors should ask corresponding authors todeclare whether they had assistance with study design, datacollection, data analysis, or manuscript preparation. If suchassistance was available, the authors should disclose theidentity of the individuals who provided this assistance andthe entity that supported it in the published article. Finan-cial and material support should also be acknowledged.

Groups of persons who have contributed materially tothe paper but whose contributions do not justify author-ship may be listed under such headings as “clinical inves-tigators” or “participating investigators,” and their functionor contribution should be described—for example, “servedas scientific advisors,” “critically reviewed the study pro-posal,” “collected data,” or “provided and cared for studypatients.” Because readers may infer their endorsement ofthe data and conclusions, these persons must give writtenpermission to be acknowledged.

II. B. EditorshipII. B. 1. The Role of the Editor

The editor of a journal is the person responsible for itsentire content. Owners and editors of medical journalshave a common endeavor—publication of a reliable, read-able journal produced with due respect for the stated aimsof the journal and for costs. Owners and editors, however,have different functions. Owners have the right to appointand dismiss editors and to make important business deci-sions in which editors should be involved to the fullestextent possible. Editors must have full authority for deter-mining the editorial content of the journal. The concept ofeditorial freedom should be resolutely defended by editorseven to the extent of their placing their positions at stake.To secure this freedom in practice, the editor should havedirect access to the highest level of ownership, not to adelegated manager.

Editors of medical journals should have a contract thatclearly states their rights and duties, the general termsof the appointment, and the mechanisms for resolvingconflict.

An independent editorial advisory board may be use-ful in helping the editor establish and maintain editorialpolicy.

II. B. 2. Editorial Freedom

The ICMJE adopts the World Association of MedicalEditors’ definition of editorial freedom. According to thisdefinition, editorial freedom, or independence, is the con-cept that editors-in-chief have full authority over the edi-torial content of their journal and the timing of publica-tion of that content. Journal owners should not interfere inthe evaluation, selection, or editing of individual articleseither directly or by creating an environment that stronglyinfluences decisions. Journal owners should not require ed-itors to publish supplements as part of their contractual


www.icmje.org 3

agreements. Editors should base decisions on the validity ofthe work and its importance to the journal’s readers, noton the commercial success of the journal. Editors should befree to express critical but responsible views about all as-pects of medicine without fear of retribution, even if theseviews conflict with the commercial goals of the publisher.Editors and editors’ organizations are obligated to supportthe concept of editorial freedom and to draw major trans-gressions of such freedom to the attention of the interna-tional medical, academic, and lay communities.

II. C. Peer ReviewUnbiased, independent, critical assessment is an in-

trinsic part of all scholarly work, including the scientificprocess. Peer review is the critical assessment of manu-scripts submitted to journals by experts who are not part ofthe editorial staff. Peer review can therefore be viewed as animportant extension of the scientific process. Although itsactual value has been little studied and is widely debated(4), peer review helps editors decide which manuscripts aresuitable for their journals and helps authors and editors toimprove the quality of reporting. A peer-reviewed journalsubmits most of its published research articles for outsidereview. The number and kinds of manuscripts sent forreview, the number of reviewers, the reviewing procedures,and the use made of the reviewers’ opinions may vary. Inthe interests of transparency, each journal should publiclydisclose its policies and average turn-around times in itsInstructions to Authors.

II. D. Conflicts of InterestPublic trust in the peer-review process and the credi-

bility of published articles depends in part on how wellconflict of interest is handled during writing, peer review,and editorial decision making. Conflict of interest existswhen an author (or the author’s institution), reviewer, oreditor has financial or personal relationships that inappro-priately influence (bias) his or her actions (such relation-ships are also known as dual commitments, competing in-terests, or competing loyalties). These relationships varyfrom being negligible to having great potential for influ-encing judgment. Not all relationships represent true con-flict of interest. On the other hand, the potential for con-flict of interest can exist regardless of whether an individualbelieves that the relationship affects his or her scientificjudgment. Financial relationships (such as employment,consultancies, stock ownership, honoraria, and paid experttestimony) are the most easily identifiable conflicts of in-terest and the most likely to undermine the credibility ofthe journal, the authors, and of science itself. How-ever, conflicts can occur for other reasons, such as personalrelationships, academic competition, and intellectualpassion.

All participants in the peer-review and publicationprocess must disclose all relationships that could be viewedas potential conflicts of interest. Disclosure of such rela-tionships is also important in connection with editorials

and review articles, because it can be more difficult to de-tect bias in these types of publications than in reports oforiginal research. Editors may use information disclosed inconflict-of-interest and financial-interest statements as abasis for editorial decisions. Editors should publish thisinformation if they believe it is important in judging themanuscript.

II. D. 1. Potential Conflicts of Interest Related toIndividual Authors’ Commitments

When authors submit a manuscript, whether an articleor a letter, they are responsible for disclosing all financialand personal relationships that might bias their work. Toprevent ambiguity, authors must state explicitly whetherpotential conflicts do or do not exist. Authors should do soin the manuscript on a conflict-of-interest notification pagethat follows the title page, providing additional detail, ifnecessary, in a cover letter that accompanies the manu-script. (See Section IV. A. 3. Conflict-of-Interest Disclosure.)The ICMJE developed a uniform disclosure form thatICMJE member journals piloted in 2009. The second ver-sion of the form is now available, as is an accompanyingGlossary. Other journals are welcome to adopt this form.

Authors should identify individuals who provide writ-ing or other assistance and disclose the funding source forthis assistance.

Investigators must disclose potential conflicts to studyparticipants and should state in the manuscript whetherthey have done so.

Editors also need to decide whether to publish infor-mation disclosed by authors about potential conflicts. Ifdoubt exists, it is best to err on the side of publication.

II. D. 2. Potential Conflicts of Interest Related toProject Support

Increasingly, individual studies receive funding fromcommercial firms, private foundations, and government.The conditions of this funding have the potential to biasand otherwise discredit the research.

Scientists have an ethical obligation to submit credit-able research results for publication. Researchers shouldnot enter into agreements that interfere with their access toall of the data and their ability to analyze them indepen-dently, and to prepare and publish manuscripts. Authorsshould describe the role of the study sponsor, if any, instudy design; collection, analysis, and interpretation ofdata; writing the report; and the decision to submit thereport for publication. If the supporting source had nosuch involvement, the authors should so state. Biases po-tentially introduced when sponsors are directly involved inresearch are analogous to methodological biases. Somejournals, therefore, choose to include information in theMethods section about the sponsor’s involvement.

Editors may request that authors of a study funded byan agency with a proprietary or financial interest in the


4 www.icmje.org

http://www.icmje.org/coi_disclosure.pdf

http://www.icmje.org/coi_glossary.pdf

outcome sign a statement, such as “I had full access to allof the data in this study and I take complete responsibilityfor the integrity of the data and the accuracy of the dataanalysis.” Editors should be encouraged to review copies ofthe protocol and/or contracts associated with project-specific studies before accepting such studies for publica-tion. Editors may request a statistical analysis of all data byan independent biostatistician. Editors may choose not toconsider an article if a sponsor has asserted control over theauthors’ right to publish.

II. D. 3. Potential Conflicts of Interest Related toCommitments of Editors, Journal Staff, or Reviewers

Editors should avoid selecting external peer reviewerswith obvious potential conflicts of interest—for example,those who work in the same department or institution asany of the authors. Authors often provide editors with thenames of persons they feel should not be asked to review amanuscript because of potential, usually professional, con-flicts of interest. When possible, authors should be asked toexplain or justify their concerns; that information is impor-tant to editors in deciding whether to honor such requests.

Reviewers must disclose to editors any conflicts of in-terest that could bias their opinions of the manuscript, andthey should recuse themselves from reviewing specificmanuscripts if the potential for bias exists. As in the case ofauthors, silence on the part of reviewers concerning poten-tial conflicts may mean either that conflicts exist and thereviewer has failed to disclose them or conflicts do notexist. Reviewers must therefore also be asked to state ex-plicitly whether conflicts do or do not exist. Reviewersmust not use knowledge of the work, before its publica-tion, to further their own interests.

Editors who make final decisions about manuscriptsmust have no personal, professional, or financial involve-ment in any of the issues they might judge. Other mem-bers of the editorial staff, if they participate in editorialdecisions, must provide editors with a current descriptionof their financial interests (as they might relate to editorialjudgments) and recuse themselves from any decisions inwhich a conflict of interest exists. Editorial staff must notuse information gained through working with manuscriptsfor private gain. Editors should publish regular disclosurestatements about potential conflicts of interests related tothe commitments of journal staff.

II. E. Privacy and ConfidentialityII. E. 1. Patients and Study Participants

Patients have a right to privacy that should not beviolated without informed consent. Identifying informa-tion, including names, initials, or hospital numbers, shouldnot be published in written descriptions, photographs, orpedigrees unless the information is essential for scientificpurposes and the patient (or parent or guardian) gives writ-ten informed consent for publication. Informed consentfor this purpose requires that an identifiable patient be

shown the manuscript to be published. Authors shoulddisclose to these patients whether any potential identifiablematerial might be available via the Internet as well as inprint after publication. Patient consent should be writtenand archived with the journal, the authors, or both, asdictated by local regulations or laws. Applicable laws varyfrom locale to locale, and journals should establish theirown policies with legal guidance. Since a journal that ar-chives the consent will be aware of patient identity, somejournals may decide that patient confidentiality is betterguarded by having the author archive the consent and in-stead providing the journal with a written statement thatattests that they have received and archived written patientconsent.

Nonessential identifying details should be omitted. In-formed consent should be obtained if there is any doubtthat anonymity can be maintained. For example, maskingthe eye region in photographs of patients is inadequateprotection of anonymity. If identifying characteristics arealtered to protect anonymity, such as in genetic pedigrees,authors should provide assurance, and editors should sonote, that such alterations do not distort scientific meaning.

The requirement for informed consent should be in-cluded in the journal’s Instructions for Authors. When in-formed consent has been obtained, it should be indicatedin the published article.

II. E. 2. Authors and Reviewers

Manuscripts must be reviewed with due respect forauthors’ confidentiality. In submitting their manuscriptsfor review, authors entrust editors with the results of theirscientific work and creative effort, on which their reputa-tion and career may depend. Authors’ rights may be vio-lated by disclosure of the confidential details during reviewof their manuscript. Reviewers also have rights to confiden-tiality, which must be respected by the editor. Confidenti-ality may have to be breached if dishonesty or fraud isalleged but otherwise must be honored.

Editors must not disclose information about manu-scripts (including their receipt, content, status in the re-viewing process, criticism by reviewers, or ultimate fate) toanyone other than the authors and reviewers. This includesrequests to use the materials for legal proceedings.

Editors must make clear to their reviewers that manu-scripts sent for review are privileged communications andare the private property of the authors. Therefore, review-ers and members of the editorial staff must respect theauthors’ rights by not publicly discussing the authors’ workor appropriating their ideas before the manuscript is pub-lished. Reviewers must not be allowed to make copies ofthe manuscript for their files and must be prohibited fromsharing it with others, except with the editor’s permission.Reviewers should return or destroy copies of manuscriptsafter submitting reviews. Editors should not keep copies ofrejected manuscripts.

Reviewer comments should not be published or oth-


www.icmje.org 5

erwise publicized without permission of the reviewer, au-thor, and editor.

Opinions differ on whether reviewers should remainanonymous. Authors should consult the Information forAuthors of the journal to which they have chosen to sub-mit a manuscript to determine whether reviews are anon-ymous. When comments are not signed, the reviewers’identity must not be revealed to the author or anyone elsewithout the reviewers’ permission.

Some journals publish reviewers’ comments with themanuscript. No such procedure should be adopted withoutthe consent of the authors and reviewers. However, review-ers’ comments should be sent to other persons reviewingthe same manuscript, which helps reviewers learn from thereview process. Reviewers also may be notified of the edi-tor’s decision to accept or reject a manuscript.

II. F. Protection of Human Subjects and Animals inResearch

When reporting experiments on human subjects, au-thors should indicate whether the procedures followedwere in accordance with the ethical standards of the re-sponsible committee on human experimentation (institu-tional and national) and with the Helsinki Declaration of1975, as revised in 2008 (5). If doubt exists whether theresearch was conducted in accordance with the HelsinkiDeclaration, the authors must explain the rationale fortheir approach and demonstrate that the institutional re-view body explicitly approved the doubtful aspects of thestudy. When reporting experiments on animals, authorsshould indicate whether the institutional and nationalguide for the care and use of laboratory animals wasfollowed.

III. PUBLISHING AND EDITORIAL ISSUES RELATED TO

PUBLICATION IN BIOMEDICAL JOURNALS

III. A. Obligation to Publish Negative StudiesEditors should seriously consider for publication any

carefully done study of an important question, relevant totheir readers, whether the results for the primary or anyadditional outcome are statistically significant. Failure tosubmit or publish findings because of lack of statisticalsignificance is an important cause of publication bias.

III. B. Corrections, Retractions, and “Expressions ofConcern”

Editors must assume initially that authors are report-ing work based on honest observations. Nevertheless, twotypes of difficulty may arise.

First, errors may be noted in published articles thatrequire the publication of a correction or erratum on partof the work. The corrections should appear on a numberedpage, be listed in the Table of Contents, include the com-plete original citation, and link to the original article andvice versa if online. It is conceivable that an error could beso serious as to vitiate the entire body of the work, but thisis unlikely and should be addressed by editors and authors

on an individual basis. Such an error should not be con-fused with inadequacies exposed by the emergence of newscientific information in the normal course of research.The latter requires no corrections or withdrawals.

The second type of difficulty is scientific fraud. If sub-stantial doubt arises about the honesty or integrity of work,either submitted or published, it is the editor’s responsibil-ity to ensure that the question is appropriately pursued,usually by the authors’ sponsoring institution. Ordinarily,it is not the responsibility of the editor to conduct a fullinvestigation or to make a determination—that responsi-bility lies with the institution where the work was done orwith the funding agency. The editor should be promptlyinformed of the final decision, and if a fraudulent paperhas been published, the journal must print a retraction. Ifthis method of investigation does not result in a satisfac-tory conclusion, the editor may choose to conduct his orher own investigation. As an alternative to retraction, theeditor may choose to publish an expression of concernabout aspects of the conduct or integrity of the work.

The retraction or expression of concern, so labeled,should appear on a numbered page in a prominent sectionof the print journal as well as in the online version, belisted in the Table of Contents page, and include in itsheading the title of the original article. It should not simplybe a letter to the editor. Ideally, the first author of theretraction should be the same as that of the article, al-though under certain circumstances the editor may acceptretractions by other responsible persons. The text of theretraction should explain why the article is being retractedand include a complete citation reference to that article.

The validity of previous work by the author of a fraud-ulent paper cannot be assumed. Editors may ask the au-thor’s institution to assure them of the validity of earlierwork published in their journals or to retract it. If this isnot done, editors may choose to publish an announcementexpressing concern that the validity of previously publishedwork is uncertain.

Editors who have questions related to editorial or sci-entific misconduct may find it useful to consult the excel-lent flow charts that the Committee on Publication Ethics(COPE) has developed (http://www.publicationethics.org.uk). COPE, which was formed in 1997, is a forum inwhich editors of peer-reviewed journals can discuss issuesrelated to the integrity of the scientific record; it supportsand encourages editors to report, catalogue, and instigateinvestigations into ethical problems in the publication pro-cess. COPE’s major objective is to provide a sounding boardfor editors struggling with how best to deal with possiblebreaches in research and publication ethics.

III. C. CopyrightMany biomedical journals ask authors to transfer

copyright to the journal. However, an increasing numberof “open-access” journals do not require transfer of copy-right. Editors should make their position on copyright


6 www.icmje.org

transfer clear to authors and to others who might be inter-ested in using editorial content from their journals. Thecopyright status of articles in a given journal can vary:Some content cannot be copyrighted (for example, articleswritten by employees of the U.S. or some other govern-ments in the course of their work); editors may agree towaive copyright on others; and still others may be protectedunder serial rights (that is, use in publications other than jour-nals, including electronic publications, is permitted).

III. D. Overlapping PublicationsIII. D. 1. Duplicate Submission

Most biomedical journals will not consider manu-scripts that are simultaneously being considered by otherjournals. Among the principal considerations that have ledto this policy are: 1) the potential for disagreement whentwo (or more) journals claim the right to publish a manu-script that has been submitted simultaneously to more thanone; and 2) the possibility that two or more journals willunknowingly and unnecessarily undertake the work of peerreview, edit the same manuscript, and publish the samearticle.

However, editors of different journals may decide tosimultaneously or jointly publish an article if they believethat doing so would be in the best interest of public health.

III. D. 2. Redundant Publication

Redundant (or duplicate) publication is publication ofa paper that overlaps substantially with one already pub-lished in print or electronic media.

Readers of primary source periodicals, whether printor electronic, deserve to be able to trust that what they arereading is original unless there is a clear statement that theauthor and editor are intentionally republishing an article.The bases of this position are international copyright laws,ethical conduct, and cost-effective use of resources. Dupli-cate publication of original research is particularly prob-lematic because it can result in inadvertent double-counting or inappropriate weighting of the results of asingle study, which distorts the available evidence.

Most journals do not wish to receive papers on workthat has already been reported in large part in a publishedarticle or is contained in another paper that has been sub-mitted or accepted for publication elsewhere, in print or inelectronic media. This policy does not preclude the journalfrom considering a paper that has been rejected by anotherjournal, or a complete report that follows publication of apreliminary report, such as an abstract or poster displayedat a professional meeting. It also does not prevent journalsfrom considering a paper that has been presented at a sci-entific meeting but was not published in full, or that isbeing considered for publication in a proceedings or simi-lar format. Brief press reports of scheduled meetings arenot usually regarded as breaches of this rule, but they maybe if additional data or copies of tables and figures amplifysuch reports. The ICMJE does not consider results posted

in clinical trial registries as previous publication if the re-sults are presented in the same, ICMJE-accepted registry inwhich initial registration of trial methods occurred and ifthe results are posted in the form of a brief structuredabstract or table. The ICMJE also believes that the resultsregistry should either cite full publications of the resultswhen available or include a statement that indicates thatthe results have not yet been published in a peer-reviewedjournal.

When submitting a paper, the author must alwaysmake a complete statement to the editor about all submis-sions and previous reports (including meeting presenta-tions and posting of results in registries) that might beregarded as redundant or duplicate publication. The au-thor must alert the editor if the manuscript includes sub-jects about which the authors have published a previousreport or have submitted a related report to another pub-lication. Any such report must be referred to and refer-enced in the new paper. Copies of such material should beincluded with the submitted manuscript to help the editordecide how to handle the matter.

If redundant or duplicate publication is attempted oroccurs without such notification, authors should expect ed-itorial action to be taken. At the least, prompt rejection ofthe submitted manuscript should be expected. If the editorwas not aware of the violations and the article has alreadybeen published, then a notice of redundant or duplicatepublication will probably be published with or without theauthor’s explanation or approval.

Preliminary reporting to public media, governmentalagencies, or manufacturers of scientific information de-scribed in a paper or a letter to the editor that has beenaccepted but not yet published violates the policies ofmany journals. Such reporting may be warranted when thepaper or letter describes major therapeutic advances orpublic health hazards, such as serious adverse effects ofdrugs, vaccines, other biological products, medicinal de-vices, or reportable diseases. This reporting should notjeopardize publication, but should be discussed with andagreed upon by the editor in advance.

III. D. 3. Acceptable Secondary Publication

Certain types of articles, such as guidelines producedby governmental agencies and professional organizations,may need to reach the widest possible audience. In suchinstances, editors sometimes deliberately publish materialthat is also being published in other journals, with theagreement of the authors and the editors of those journals.Secondary publication for various other reasons, in thesame or another language, especially in other countries, isjustifiable and can be beneficial provided that the followingconditions are met.

1. The authors have received approval from the editorsof both journals (the editor concerned with secondary pub-lication must have a photocopy, reprint, or manuscript ofthe primary version).


www.icmje.org 7

2. The priority of the primary publication is respectedby a publication interval of at least 1 week (unless specifi-cally negotiated otherwise by both editors).

3. The paper for secondary publication is intended fora different group of readers; an abbreviated version couldbe sufficient.

4. The secondary version faithfully reflects the dataand interpretations of the primary version.

5. The footnote on the title page of the secondaryversion informs readers, peers, and documenting agenciesthat the paper has been published in whole or in part andstates the primary reference. A suitable footnote mightread: “This article is based on a study first reported in the[title of journal, with full reference].”

Permission for such secondary publication should befree of charge.

6. The title of the secondary publication should indi-cate that it is a secondary publication (complete republica-tion, abridged republication, complete translation, orabridged translation) of a primary publication. Of note,the NLM does not consider translations to be “republica-tions” and does not cite or index translations when theoriginal article was published in a journal that is indexed inMEDLINE.

7. Editors of journals that simultaneously publish inmultiple languages should understand that NLM indexesthe primary language version. When the full text of anarticle appears in more than one language in a journal issue(such as Canadian journals with the article in both Englishand French), both languages are indicated in the MED-LINE citation (for example, Mercer K. The relentless chal-lenge in health care. Healthc Manage Forum. 2008Summer;21(2):4-5. English, French. No abstract available.PMID:18795553.)

III. D. 4. Competing Manuscripts Based on the Same Study

Publication of manuscripts to air the disputes of co-investigators may waste journal space and confuse readers.On the other hand, if editors knowingly publish a manu-script written by only some of a collaborating team, theycould be denying the rest of the team their legitimate co-authorship rights and journal readers access to legitimatedifferences of opinion about the interpretation of a study.

Two kinds of competing submissions are considered:submissions by coworkers who disagree on the analysis andinterpretation of their study, and submissions by coworkerswho disagree on what the facts are and which data shouldbe reported.

Setting aside the unresolved question of ownership ofthe data, the following general observations may help edi-tors and others address such problems.

III. D. 4. a. Differences in Analysis or Interpretation

If the dispute centers on the analysis or interpretationof data, the authors should submit a manuscript that

clearly presents both versions. The difference of opinionshould be explained in a cover letter. The normal processof peer and editorial review may help the authors to resolvetheir disagreement regarding analysis or interpretation.

If the dispute cannot be resolved and the study meritspublication, both versions should be published. Optionsinclude publishing two papers on the same study, or asingle paper with two analyses or interpretations. In suchcases, it would be appropriate for the editor to publish astatement outlining the disagreement and the journal’s in-volvement in attempts to resolve it.

III. D. 4. b. Differences in Reported Methods or Results

If the dispute centers on differing opinions of whatwas actually done or observed during the study, the journaleditor should refuse publication until the disagreement isresolved. Peer review cannot be expected to resolve suchproblems. If there are allegations of dishonesty or fraud,editors should inform the appropriate authorities; authorsshould be notified of an editor’s intention to report a sus-picion of research misconduct.

III. D. 5. Competing Manuscripts Based on the Same Database

Editors sometimes receive manuscripts from separateresearch groups that have analyzed the same data set (forexample, from a public database). The manuscripts maydiffer in their analytic methods, conclusions, or both. Eachmanuscript should be considered separately. If interpreta-tion of the data is very similar, it is reasonable but notmandatory for editors to give preference to the manuscriptthat was received first. However, editorial consideration ofmultiple submissions may be justified under these circum-stances, and there may even be a good reason to publishmore than one manuscript because different analytical ap-proaches may be complementary and equally valid.

III. E. CorrespondenceThe corresponding author/guarantor has primary re-

sponsibility for correspondence with the journal, but theICMJE recommends that editors send a copy of any cor-respondence to all listed authors.

Biomedical journals should provide the readershipwith a mechanism for submitting comments, questions, orcriticisms about published articles, as well as brief reportsand commentary unrelated to previously published articles.This probably but not necessarily takes the form of a cor-respondence section or column. The authors of articlesdiscussed in correspondence should be given an opportu-nity to respond, preferably in the same issue in which theoriginal correspondence appears. Authors of correspon-dence should be asked to declare any competing or con-flicting interests.

Published correspondence may be edited for length,grammatical correctness, and journal style. Alternatively,editors may choose to publish unedited correspondence,for example in rapid-response sections on the Internet. The


8 www.icmje.org

journal should declare its editorial practices in this regard.Authors should approve editorial changes that alter thesubstance or tone of a letter or response. In all instances,editors must make an effort to screen discourteous, inaccu-rate, or libelous statements and should not allow ad hominemarguments intended to discredit opinions or findings.

Although editors have the prerogative to reject corre-spondence that is irrelevant, uninteresting, or lacking co-gency, they have a responsibility to allow a range of opin-ions to be expressed. The correspondence column shouldnot be used merely to promote the journal’s or the editors’point of view.

In the interests of fairness and to keep correspondencewithin manageable proportions, journals may want to settime limits for responding to published material and fordebate on a given topic. Journals should also decidewhether they would notify authors when correspondencebearing on their published work is going to appear in stan-dard or rapid-response sections. Journals should also setpolicy with regard to the archiving of unedited correspon-dence that appears online. These policies should be pub-lished both in print and electronic versions of the journal.

III. F. Supplements, Theme Issues, and Special SeriesSupplements are collections of papers that deal with

related issues or topics, are published as a separate issue ofthe journal or as part of a regular issue, and are usuallyfunded by sources other than the journal’s publisher. Thereis evidence that supplement content can be of lower qualitythan the content of the parent journal (6). Because fundingsources can bias the content of supplements through thechoice of topics and viewpoints, journals should consideradopting the following principles. These same principlesapply to theme issues or special series that have externalfunding and/or guest editors.

1. The journal editor must be given and take full re-sponsibility for the policies, practices, and content of sup-plements, including complete control of the decision toselect authors, peer reviewers, and content for the supple-ment. Editing by the funding organization should not bepermitted.

2. The journal editor must retain the authority to sendsupplement manuscripts for external peer review and toreject manuscripts submitted for the supplement. Theseconditions should be made known to authors and externalsupplement editors before beginning editorial work on thesupplement.

3. The journal editor must approve the appointmentof any external editor of the supplement and take respon-sibility for the work of the external editor.

4. The source of the idea for the supplement, sourcesof funding for the research, publication, and products ofthe funding source that are considered in the supplementshould be clearly stated and prominently located in the

supplement, preferably on each page. Whenever possible,supplements should be funded by more than one sponsor.

5. Advertising in supplements should follow the samepolicies as those of the rest of the journal.

6. Journal editors must enable readers to distinguishreadily between ordinary editorial pages and supplementpages.

7. Journal editors and supplement editors must notaccept personal favors or remuneration from sponsors ofsupplements.

8. Secondary publication in supplements (republica-tion of papers published elsewhere) should be clearly iden-tified by the citation of the original paper. Supplementsshould avoid redundant or duplicate publication. Supple-ments should not republish research results, but republica-tion of guidelines or other material in the public interestmight be appropriate.

9. The principles of authorship and disclosure of po-tential conflicts of interest discussed elsewhere in this doc-ument should be applied to supplements.

III. G. Electronic PublishingMost biomedical journals are now published in elec-

tronic as well as print versions, and some are publishedonly in electronic form. Because electronic publishing(which includes the Internet) is the same as publishing inprint, in the interests of clarity and consistency the recom-mendations of this document should be applied to elec-tronically published medical and health information.

The nature of electronic publication requires somespecial considerations, both within and beyond this docu-ment. At a minimum, Web sites should indicate the fol-lowing: names, appropriate credentials, affiliations, and rel-evant conflicts of interest of editors, authors, andcontributors; documentation and attribution of referencesand sources for all content; information about copyright;disclosure of site ownership; and disclosure of sponsorship,advertising, and commercial funding.

Linking from one health or medical Internet site toanother may be perceived as an implicit recommendationof the quality of the second site. Journals thus should ex-ercise caution in linking to other sites; when users are link-ing to another site, it may be helpful to provide an explicitstatement that they are leaving the journal’s site. Links toother sites posted as a result of financial considerationsshould be clearly indicated as such. All dates of contentposting and updating should be indicated. In electroniclayout as in print, advertising and promotional messagesshould not be juxtaposed with editorial content, and com-mercial content should be clearly identified as such.

Electronic publication is in flux. Editors should de-velop, make available to authors, and implement policieson issues unique to electronic publishing. These issues in-clude archiving, error correction, version control, choice ofthe electronic or print version of the journal as the journalof record, and publication of ancillary material.


www.icmje.org 9

Under no circumstances should a journal remove anarticle from its Web site or archive. If a correction or re-traction becomes necessary, the explanation must be la-beled appropriately and communicated as soon as possibleon a citable page in a subsequent issue of the journal.

Preservation of electronic articles in a permanent ar-chive is essential for the historical record. Access to thearchive should be immediate and controlled by a thirdparty, such as a library, instead of the publisher. Deposi-tion in multiple archives is encouraged.

III. H. AdvertisingMost medical journals carry advertising, which gener-

ates income for their publishers, but advertising must notbe allowed to influence editorial decisions. Journals shouldhave formal, explicit, written policies for advertising inboth print and electronic versions; Web site advertisingpolicy should parallel that for the printed journals. Editorsmust have full and final authority for approving advertise-ments and enforcing advertising policy.

When possible, editors should make use of the judg-ments of independent bodies for reviewing advertising.Readers should be able to distinguish readily between ad-vertising and editorial material. The juxtaposition of edi-torial and advertising material on the same products orsubjects should be avoided. Interspersing advertising pageswithin articles interrupts the flow of editorial content andshould be discouraged. Advertising should not be sold onthe condition that it will appear in the same issue as aparticular article.

Journals should not be dominated by advertising, buteditors should be careful about publishing advertisementsfrom only one or two advertisers, as readers may perceivethat these advertisers have influenced the editor.

Journals should not carry advertisements for productsthat have proved to be seriously harmful to health—forexample, tobacco. Editors should ensure that existing reg-ulatory or industry standards for advertisements specific totheir country are enforced, or develop their own standards.The interests of organizations or agencies should not con-trol classified and other nondisplay advertising, exceptwhere required by law. Finally, editors should consider allcriticisms of advertisements for publication.

III. I. Medical Journals and the General MediaThe public’s interest in news of medical research has

led the popular media to compete vigorously for informa-tion about research. Researchers and institutions some-times encourage reporting research in the nonmedical me-dia before full publication in a scientific journal by holdinga press conference or giving interviews.

The public is entitled to important medical informa-tion within a reasonable amount of time, and editors havea responsibility to facilitate the process. Biomedical jour-nals are published primarily for their readers, but the gen-eral public has a legitimate interest in their content: An

appropriate balance between these considerations shouldguide the journal’s interaction with the media. Doctors inpractice need to have reports available in full detail beforethey can advise their patients about the reports’ conclu-sions. Moreover, media reports of scientific research beforethe work has been peer-reviewed and fully vetted may leadto dissemination of inaccurate or premature conclusions.

An embargo system has been established in somecountries to prevent publication of stories in the generalmedia before publication of the original research in thejournal. The embargo creates a “level playing field,” whichmost reporters appreciate since it minimizes the pressureon them to publish stories they have not had time to pre-pare carefully. Consistency in the timing of public releaseof biomedical information is also important in minimizingeconomic chaos, since some articles contain informationthat has great potential to influence financial markets. Onthe other hand, the embargo system has been challenged asbeing self-serving of journals’ interests and an impedimentto rapid dissemination of scientific information.

Editors may find the following recommendations use-ful as they seek to establish policies on these issues.

● Editors can foster the orderly transmission of med-ical information from researchers, through peer-reviewedjournals, to the public. This can be accomplished by anagreement with authors that they will not publicize theirwork while their manuscript is under consideration orawaiting publication and an agreement with the media thatthey will not release stories before publication of the orig-inal research in the journal, in return for which the journalwill cooperate with them in preparing accurate stories.

● Editors need to keep in mind that an embargo sys-tem works on the honor system; no formal enforcement orpolicing mechanism exists. The decision of a signifi-cant number of media outlets or biomedical journals not torespect the embargo system would lead to its rapiddissolution.

● Very little medical research has such clear and ur-gently important clinical implications for the public’shealth that the news must be released before full publica-tion in a journal. However, if such exceptional circum-stances occur, the appropriate authorities responsible forpublic health should decide whether to disseminate infor-mation to physicians and the media in advance and shouldbe responsible for this decision. If the author and the ap-propriate authorities wish to have a manuscript consideredby a particular journal, the editor should be consulted be-fore any public release. If editors acknowledge the need forimmediate release, they should waive their policies limitingprepublication publicity.

● Policies designed to limit prepublication publicityshould not apply to accounts in the media of presentationsat scientific meetings or to the abstracts from these meet-ings (see Redundant Publication). Researchers who presenttheir work at a scientific meeting should feel free to discusstheir presentations with reporters, but they should be dis-


10 www.icmje.org

couraged from offering more detail about their study thanwas presented in the talk.

● When an article is soon to be published, editorsshould help the media prepare accurate reports by provid-ing news releases, answering questions, supplying advancecopies of the journal, or referring reporters to the appro-priate experts. This assistance should be contingent on themedia’s cooperation in timing the release of a story tocoincide with publication of the article.

● Editors, authors, and the media should apply theabove-stated principles to material released early in elec-tronic versions of journals.

III. J. Obligation to Register Clinical TrialsThe ICMJE believes that it is important to foster a

comprehensive, publicly available database of clinical trials.The ICMJE defines a clinical trial as any research projectthat prospectively assigns human subjects to interventionor concurrent comparison or control groups to study thecause-and-effect relationship between a medical interven-tion and a health outcome. Medical interventions includedrugs, surgical procedures, devices, behavioral treatments,process-of-care changes, and the like.

The ICMJE member journals will require, as a condi-tion of consideration for publication in their journals, reg-istration in a public trials registry. The details of this policyare contained in a series of editorials (see Editorials, underFrequently Asked Questions). The ICMJE encourages ed-itors of other biomedical journals to adopt similar policy.

The ICMJE does not advocate one particular registry,but its member journals will require authors to registertheir trial in a registry that meets several criteria. The reg-istry must be accessible to the public at no charge. It mustbe open to all prospective registrants and managed by anot-for-profit organization. There must be a mechanism toensure the validity of the registration data, and the registryshould be electronically searchable. Trial registration withmissing fields or fields that contain uninformative termi-nology is inadequate.

It is important to note that the ICMJE requires regis-tration of trial methodology but does not require registra-tion of trial results; it recognizes the potential problemsthat could arise from the posting of research results thathave not been subjected to an independent peer-reviewprocess. However, the ICMJE understands that the U.S.Food and Drug Administration Amendments Act of 2007(FDAAA) does require researchers to register results. TheICMJE will not consider results to be previous publicationif they are posted in the same primary clinical trial registryas the initial registration and if the results are posted in thetabular form dictated by the FDAAA. Researchers shouldbe aware that editors of journals that follow the ICMJErecommendations may consider more detailed descriptionof trial results and results published in registries other thanthe primary registry (in the case of FDAAA, ClinicalTrials-.gov) to be prior publication. The ICMJE anticipates that

the climate for results registration will change dramaticallyover coming years and the ICMJE may need to amendthese recommendations as additional agencies instituteother mandates related to results registration.

The ICMJE recommends that journals publish thetrial registration number at the end of the abstract. TheICMJE also recommends that, whenever a registrationnumber is available, authors list this number the first timethey use a trial acronym to refer to either the trial theyare reporting or to other trials that they mention in themanuscript.

IV. MANUSCRIPT PREPARATION AND SUBMISSION

IV. A. Preparing a Manuscript for Submission to aBiomedical Journal

Editors and reviewers spend many hours readingmanuscripts, and therefore appreciate receiving manu-scripts that are easy to read and edit. Much of the infor-mation in a journal’s Instructions to Authors is designed toaccomplish that goal in ways that meet each journal’s par-ticular editorial needs. The following information providesguidance in preparing manuscripts for any journal.

IV. A. 1. a. General Principles

The text of observational and experimental articles isusually (but not necessarily) divided into the following sec-tions: Introduction, Methods, Results, and Discussion.This so-called “IMRAD” structure is not an arbitrary pub-lication format but rather a direct reflection of the processof scientific discovery. Long articles may need subheadingswithin some sections (especially Results and Discussion) toclarify their content. Other types of articles, such as casereports, reviews, and editorials, probably need to be for-matted differently.

Electronic formats have created opportunities for add-ing details or whole sections, layering information, cross-linking or extracting portions of articles, and the like onlyin the electronic version. Authors need to work closely witheditors in developing or using such new publication for-mats and should submit supplementary electronic materialfor peer review.

Double-spacing all portions of the manuscript—in-cluding the title page, abstract, text, acknowledgments, ref-erences, individual tables, and legends—and generous mar-gins make it possible for editors and reviewers to edit thetext line by line and add comments and queries directly onthe paper copy. If manuscripts are submitted electronically,the files should be double-spaced to facilitate printing forreviewing and editing.

Authors should number all of the pages of the manu-script consecutively, beginning with the title page, to facil-itate the editorial process.


www.icmje.org 11

IV. A. 1. b. Reporting Guidelines for Specific StudyDesigns

Research reports frequently omit important informa-tion. Reporting guidelines have been developed for a num-ber of study designs that some journals may ask authorsto follow. Authors should consult the Information forAuthors of the journal they have chosen.

The general requirements listed in the next sectionrelate to reporting essential elements for all study designs.Authors are encouraged also to consult reporting guidelinesrelevant to their specific research design. A good source ofreporting guidelines is the EQUATOR Network (http://www.equator-network.org/home/).

IV. A .2. Title Page

The title page should have the following information:1. Article title. Concise titles are easier to read than

long, convoluted ones. Titles that are too short may, how-ever, lack important information, such as study design(which is particularly important in identifying randomized,controlled trials). Authors should include all informationin the title that will make electronic retrieval of the articleboth sensitive and specific.

2. Authors’ names and institutional affiliations. Somejournals publish each author’s highest academic degree(s),while others do not.

3. The name of the department(s) and institution(s)to which the work should be attributed.

4. Disclaimers, if any.5. Contact information for corresponding authors.

The name, mailing address, telephone and fax numbers,and e-mail address of the author responsible for correspon-dence about the manuscript (the “corresponding author;”this author may or may not be the “guarantor” for theintegrity of the study). The corresponding author shouldindicate clearly whether his or her e-mail address can bepublished.

6. The name and address of the author to whom re-quests for reprints should be addressed or a statement thatreprints are not available from the authors.

7. Source(s) of support in the form of grants, equip-ment, drugs, or all of these.

8. A running head. Some journals request a short run-ning head or footline, usually no more than 40 characters(including letters and spaces) at the foot of the title page.Running heads are published in most journals, but are alsosometimes used within the editorial office for filing andlocating manuscripts.

9. Word counts. A word count for the text only (ex-cluding abstract, acknowledgments, figure legends, and ref-erences) allows editors and reviewers to assess whether theinformation contained in the paper warrants the amount ofspace devoted to it, and whether the submitted manuscriptfits within the journal’s word limits. A separate word countfor the Abstract is useful for the same reason.

10. The number of figures and tables. It is difficult foreditorial staff and reviewers to determine whether the fig-ures and tables that should have accompanied a manuscriptwere actually included unless the numbers of figures andtables are noted on the title page.

IV. A. 3. Conflict-of-Interest Notification Page

To prevent potential conflicts of interest from beingoverlooked or misplaced, this information needs to be partof the manuscript. The ICMJE has developed a uniformdisclosure form for use by ICMJE member journals(http://www.icmje.org/coi_disclosure.pdf). Other journalsare welcome to adopt this form. Individual journals maydiffer in where they include this information, and somejournals do not send information on conflicts of interest toreviewers. (See Section II. D. Conflicts of Interest.

IV. A. 4. Abstract

Structured abstracts are preferred for original researchand systematic reviews. The abstract should provide thecontext or background for the study and should state thestudy’s purpose, basic procedures (selection of study sub-jects or laboratory animals, observational and analyticalmethods), main findings (giving specific effect sizes andtheir statistical significance, if possible), principal conclu-sions, and funding sources. It should emphasize new andimportant aspects of the study or observations. Articles onclinical trials should contain abstracts that include theitems that the CONSORT group has identified as essential(http://www.consort-statement.org/? �1190 ).

Because abstracts are the only substantive portion ofthe article indexed in many electronic databases, and theonly portion many readers read, authors need to be carefulthat they accurately reflect the content of the article. Un-fortunately, the information contained in many abstractsdiffers from that in the text (7). The format required forstructured abstracts differs from journal to journal, andsome journals use more than one format; authors need toprepare their abstracts in the format specified by the jour-nal they have chosen.

The ICMJE recommends that journals publish thetrial registration number at the end of the abstract. TheICMJE also recommends that, whenever a registrationnumber is available, authors list that number the first timethey use a trial acronym to refer to either the trial they arereporting or to other trials that they mention in themanuscript.

IV. A. 5. Introduction

Provide a context or background for the study (that is,the nature of the problem and its significance). State thespecific purpose or research objective of, or hypothesistested by, the study or observation; the research objective isoften more sharply focused when stated as a question. Boththe main and secondary objectives should be clear, and any


12 www.icmje.org

prespecified subgroup analyses should be described. Pro-vide only directly pertinent references, and do not includedata or conclusions from the work being reported.

IV. A. 6. Methods

The Methods section should include only informationthat was available at the time the plan or protocol for thestudy was being written; all information obtained duringthe study belongs in the Results section.

IV. A. 6. a. Selection and Description of Participants

Describe your selection of the observational or exper-imental participants (patients or laboratory animals, in-cluding controls) clearly, including eligibility and exclusioncriteria and a description of the source population. Becausethe relevance of such variables as age and sex to the objectof research is not always clear, authors should explain theiruse when they are included in a study report—for example,authors should explain why only participants of certainages were included or why women were excluded. Theguiding principle should be clarity about how and why astudy was done in a particular way. When authors use suchvariables as race or ethnicity, they should define how theymeasured these variables and justify their relevance.

IV. A. 6. b. Technical Information

Identify the methods, apparatus (give the manufactur-er’s name and address in parentheses), and procedures insufficient detail to allow others to reproduce the results.Give references to established methods, including statisticalmethods (see below); provide references and brief descrip-tions for methods that have been published but are notwell-known; describe new or substantially modified meth-ods, give the reasons for using them, and evaluate theirlimitations. Identify precisely all drugs and chemicalsused, including generic name(s), dose(s), and route(s) ofadministration.

Authors submitting review manuscripts should includea section describing the methods used for locating, select-ing, extracting, and synthesizing data. These methodsshould also be summarized in the abstract.

IV. A. 6. c. Statistics

Describe statistical methods with enough detail to en-able a knowledgeable reader with access to the original datato verify the reported results. When possible, quantify find-ings and present them with appropriate indicators of mea-surement error or uncertainty (such as confidence inter-vals). Avoid relying solely on statistical hypothesis testing,such as P values, which fail to convey important informa-tion about effect size. References for the design of the studyand statistical methods should be to standard works whenpossible (with pages stated). Define statistical terms, abbre-viations, and most symbols. Specify the computer softwareused.

IV. A. 7. Results

Present your results in logical sequence in the text,tables, and illustrations, giving the main or most importantfindings first. Do not repeat all the data in the tables orillustrations in the text; emphasize or summarize only themost important observations. Extra or supplementary ma-terials and technical detail can be placed in an appendixwhere they will be accessible but will not interrupt the flowof the text, or they can be published solely in the electronicversion of the journal.

When data are summarized in the Results section, givenumeric results not only as derivatives (for example, per-centages) but also as the absolute numbers from which thederivatives were calculated, and specify the statistical meth-ods used to analyze them. Restrict tables and figures tothose needed to explain the argument of the paper and toassess supporting data. Use graphs as an alternative to ta-bles with many entries; do not duplicate data in graphs andtables. Avoid nontechnical uses of technical terms in statis-tics, such as “random” (which implies a randomizing de-vice), “normal,” “significant,” “correlations,” and “sample.”

Where scientifically appropriate, analyses of the databy such variables as age and sex should be included.

IV. A. 8. Discussion

Emphasize the new and important aspects of the studyand the conclusions that follow from them in the contextof the totality of the best available evidence. Do not repeatin detail data or other information given in the Introduc-tion or the Results section. For experimental studies, it isuseful to begin the discussion by briefly summarizing themain findings, then explore possible mechanisms or expla-nations for these findings, compare and contrast the resultswith other relevant studies, state the limitations of thestudy, and explore the implications of the findings for fu-ture research and for clinical practice.

Link the conclusions with the goals of the study butavoid unqualified statements and conclusions not ade-quately supported by the data. In particular, avoid makingstatements on economic benefits and costs unless themanuscript includes the appropriate economic data andanalyses. Avoid claiming priority or alluding to work thathas not been completed. State new hypotheses when war-ranted, but label them clearly as such.

IV. A. 9. References

IV. A. 9. a. General Considerations Related toReferences

Although references to review articles can be an effi-cient way to guide readers to a body of literature, reviewarticles do not always reflect original work accurately.Readers should therefore be provided with direct referencesto original research sources whenever possible. On theother hand, extensive lists of references to original work on


www.icmje.org 13

a topic can use excessive space on the printed page.Small numbers of references to key original papers oftenserve as well as more exhaustive lists, particularly sincereferences can now be added to the electronic versionof published papers, and since electronic literaturesearching allows readers to retrieve published literatureefficiently.

Avoid using abstracts as references. References to pa-pers accepted but not yet published should be designatedas “in press” or “forthcoming”; authors should obtain writ-ten permission to cite such papers as well as verificationthat they have been accepted for publication. Informationfrom manuscripts submitted but not accepted should becited in the text as “unpublished observations” with writtenpermission from the source.

Avoid citing a “personal communication” unless itprovides essential information not available from a publicsource, in which case the name of the person and date ofcommunication should be cited in parentheses in the text.For scientific articles, obtain written permission and con-firmation of accuracy from the source of a personalcommunication.

Some but not all journals check the accuracy of allreference citations; thus, citation errors sometimes appearin the published version of articles. To minimize such er-rors, references should be verified using either an electronicbibliographic source, such as PubMed or print copies fromoriginal sources. Authors are responsible for checking thatnone of the references cite retracted articles except in thecontext of referring to the retraction. For articles publishedin journals indexed in MEDLINE, the ICMJE considersPubMed the authoritative source for information aboutretractions. Authors can identify retracted articles in MED-LINE by using the following search term, where pt insquare brackets stands for publication type: Retracted pub-lication [pt] in PubMed.

IV. A. 9. b. Reference Style and Format

The Uniform Requirements style for references isbased largely on an American National Standards Institutestyle adapted by the NLM for its databases. Authors shouldconsult NLM’s Citing Medicine for information on itsrecommended formats for a variety of reference types. Au-thors may also consult sample references, a list of examplesextracted from or based on Citing Medicine for easy use bythe ICMJE audience; these sample references are main-tained by NLM.

References should be numbered consecutively in theorder in which they are first mentioned in the text. Identifyreferences in text, tables, and legends by Arabic numeralsin parentheses. References cited only in tables or figurelegends should be numbered in accordance with the se-quence established by the first identification in the text ofthe particular table or figure. The titles of journals shouldbe abbreviated according to the style used in the list of

Journals Indexed for MEDLINE, posted by the NLM onthe Library’s Web site. Journals vary on whether they askauthors to cite electronic references within parentheses inthe text or in numbered references following the text. Au-thors should consult with the journal to which they plan tosubmit their work.

IV. A. 10. Tables

Tables capture information concisely and display itefficiently; they also provide information at any desiredlevel of detail and precision. Including data in tables ratherthan text frequently makes it possible to reduce the lengthof the text.

Type or print each table with double-spacing on aseparate sheet of paper. Number tables consecutively in theorder of their first citation in the text and supply a brieftitle for each. Do not use internal horizontal or verticallines. Give each column a short or an abbreviated heading.Authors should place explanatory matter in footnotes, notin the heading. Explain all nonstandard abbreviations infootnotes, and use the following symbols, in sequence:

*, †, ‡, §, �, ¶, **, ††, ‡‡, §§, � �, ¶¶, etc.Identify statistical measures of variations, such as stan-

dard deviation and standard error of the mean.Be sure that each table is cited in the text.If you use data from another published or unpublished

source, obtain permission and acknowledge that sourcefully.

Additional tables containing backup data too extensiveto publish in print may be appropriate for publication inthe electronic version of the journal, deposited with anarchival service, or made available to readers directly by theauthors. An appropriate statement should be added to thetext to inform readers that this additional information isavailable and where it is located. Submit such tables forconsideration with the paper so that they will be availableto the peer reviewers.

IV. A. 11. Illustrations (Figures)

Figures should be either professionally drawn and pho-tographed, or submitted as photographic-quality digitalprints. In addition to requiring a version of the figuressuitable for printing, some journals now ask authors forelectronic files of figures in a format (for example, JPEG orGIF) that will produce high-quality images in the Webversion of the journal; authors should review the images ofsuch files on a computer screen before submitting them tobe sure they meet their own quality standards.

For x-ray films, scans, and other diagnostic images, aswell as pictures of pathology specimens or photomicro-graphs, send sharp, glossy, black-and-white or color pho-tographic prints, usually 127 � 173 mm (5 � 7 inches).Although some journals redraw figures, many do not. Let-ters, numbers, and symbols on figures should therefore beclear and consistent throughout, and large enough to re-


14 www.icmje.org

main legible when the figure is reduced for publication.Figures should be made as self-explanatory as possible,since many will be used directly in slide presentations. Ti-tles and detailed explanations belong in the legends—noton the illustrations themselves.

Photomicrographs should have internal scale markers.Symbols, arrows, or letters used in photomicrographsshould contrast with the background.

Photographs of potentially identifiable people must beaccompanied by written permission to use the photograph.

Figures should be numbered consecutively accordingto the order in which they have been cited in the text. If afigure has been published previously, acknowledge theoriginal source and submit written permission from thecopyright holder to reproduce the figure. Permission is re-quired irrespective of authorship or publisher except fordocuments in the public domain.

For illustrations in color, ascertain whether the journalrequires color negatives, positive transparencies, or colorprints. Accompanying drawings marked to indicate the re-gion to be reproduced might be useful to the editor. Somejournals publish illustrations in color only if the authorpays the additional cost.

Authors should consult the journal about require-ments for figures submitted in electronic formats.

IV. A. 12. Legends for Illustrations (Figures)

Type or print out legends for illustrations using dou-ble spacing, starting on a separate page, with Arabic nu-merals corresponding to the illustrations. When symbols,arrows, numbers, or letters are used to identify parts of theillustrations, identify and explain each one clearly in thelegend. Explain the internal scale and identify the methodof staining in photomicrographs.

IV. A. 13. Units of Measurement

Measurements of length, height, weight, and volumeshould be reported in metric units (meter, kilogram, orliter) or their decimal multiples.

Temperatures should be in degrees Celsius. Bloodpressures should be in millimeters of mercury, unless otherunits are specifically required by the journal.

Journals vary in the units they use for reporting hema-tologic, clinical chemistry, and other measurements. Au-thors must consult the Information for Authors of the par-ticular journal and should report laboratory information inboth local and International System of Units (SI). Editorsmay request that authors add alternative or non-SI units,since SI units are not universally used. Drug concentra-tions may be reported in either SI or mass units, but thealternative should be provided in parentheses whereappropriate.

IV. A. 14. Abbreviations and Symbols

Use only standard abbreviations; use of nonstandardabbreviations can be confusing to readers. Avoid abbrevia-tions in the title of the manuscript. The spelled-out abbre-viation followed by the abbreviation in parenthesis shouldbe used on first mention unless the abbreviation is a stan-dard unit of measurement.

IV. B. Sending the Manuscript to the Journal

An increasing number of journals now accept elec-tronic submission of manuscripts, whether on disk, as ane-mail attachment, or by downloading directly onto thejournal’s Web site. Electronic submission saves time andmoney and allows the manuscript to be handled in elec-tronic form throughout the editorial process (for example,when it is sent out for review). For specific instructions onelectronic submission, authors should consult the journal’sInstructions for Authors.

If a paper version of the manuscript is submitted, sendthe required number of copies of the manuscript and fig-ures; they are all needed for peer review and editing, andthe editorial office staff cannot be expected to make therequired copies.

Manuscripts must be accompanied by a cover letter,which should include the following information.

● A full statement to the editor about all submissionsand previous reports that might be regarded as redundantpublication of the same or very similar work. Any suchwork should be referred to specifically and referenced inthe new paper. Copies of such material should be includedwith the submitted paper to help the editor address thesituation.

● A statement of financial or other relationships thatmight lead to a conflict of interest, if that information isnot included in the manuscript itself or in an authors’form.

● A statement that the manuscript has been read andapproved by all the authors, that the requirements for au-thorship as stated earlier in this document have been met,and that each author believes that the manuscript repre-sents honest work if that information is not provided inanother form (see below).

● The name, address, and telephone number of thecorresponding author, who is responsible for communicat-ing with the other authors about revisions and final ap-proval of the proofs, if that information is not included inthe manuscript itself.

The letter should give any additional information thatmay be helpful to the editor, such as the type or format ofarticle in the particular journal that the manuscript repre-sents. If the manuscript has been submitted previously toanother journal, it is helpful to include the previous edi-tor’s and reviewers’ comments with the submitted manu-script, along with the authors’ responses to those com-ments. Editors encourage authors to submit these previous


www.icmje.org 15

communications. Doing so may expedite the reviewprocess.

Many journals now provide a presubmission checklistto help the author ensure that all the components of thesubmission have been included. Some journals now alsorequire that authors complete checklists for reports of cer-tain study types (for example, the CONSORT checklist forreports of randomized, controlled trials). Authors shouldlook to see if the journal uses such checklists, and sendthem with the manuscript if they are requested.

Letters of permission to reproduce previously pub-lished material, use previously published illustrations, re-port information about identifiable persons, or to acknowl-edge people for their contributions must accompany themanuscript.

V. REFERENCES

A. References Cited in This Document1. Davidoff F, for the CSE Task Force on Authorship.

Who’s the author? Problems with biomedical authorship,and some possible solutions. Science Editor. 2000;23:111-9.

2. Yank V, Rennie D. Disclosure of researcher contri-butions: a study of original research articles in The Lancet.Ann Intern Med. 1999;130:661-70.

3. Flanagin A, Fontanarosa PB, DeAngelis CD. Au-thorship for research groups. JAMA. 2002;288:3166-8.

4. Godlee F, Jefferson T. Peer Review in Health Sci-ences. London: BMJ Books; 1999.

5. http://www.wma.net/e/policy/b3.htm (accessedJune 26, 2009).

6. Rochon PA, Gurwitz JH, Cheung CM, Hayes JA,Chalmers TC. Evaluating the quality of articles publishedin journal supplements compared with the quality of thosepublished in the parent journal. JAMA. 1994;272:108-13.

7. Pitkin RM, Branagan MA, Burmeister LF. Accuracyof data in abstracts of published research articles. JAMA.1999;281:1110-1.

B. Other Sources of Information Related to BiomedicalJournals

World Association of Medical Editors (WAME)Council of Science Editors (CSE)European Association of Science Editors (EASE)Cochrane CollaborationCommittee on Publication Ethics (COPE)EQUATOR NETWORK http://www.equator-network

.org

VI. ABOUT THE INTERNATIONAL COMMITTEE OF

MEDICAL JOURNAL EDITORS

The ICMJE is a group of general medical journal ed-itors whose participants meet annually and fund their workon the Uniform Requirements for Manuscripts. The

ICMJE invites comments on this document and sugges-tions for agenda items.

VII. AUTHORS OF THE UNIFORM REQUIREMENTS FOR

MANUSCRIPTS SUBMITTED TO BIOMEDICAL JOURNALS

The ICMJE participating journals and organizationsand their representatives who approved the revised Uni-form Requirements for Manuscripts in April 2010 includeAnnals of Internal Medicine, British Medical Journal, Cana-dian Medical Association Journal, China Medical Journal,Croatian Medical Journal, Journal of the American MedicalAssociation, Nederlands Tijdschrift voor Geneeskunde (TheDutch Medical Journal), New England Journal of Medicine,New Zealand Medical Journal, The Lancet, The MedicalJournal of Australia, Revista Medica de Chile, Tidsskrift forDen Norske Lægeforening (The Journal of the NorwegianMedical Association), Ugeskrift for Laeger (Journal of theDanish Medical Association), the U.S. NLM, and theWorld Association of Medical Editors.

VIII. USE, DISTRIBUTION, AND TRANSLATION OF THE

UNIFORM REQUIREMENTS

Users may print, copy, and distribute this documentwithout charge for not-for-profit, educational purpose.The ICMJE does not stock paper copies (reprints) of thisdocument.

The ICMJE policy is for interested organizations tolink to the official English language document at www.ICMJE.org. The ICMJE does not endorse posting of thedocument on Web sites other than that of the ICMJE.

The ICMJE welcomes organizations to reprint ortranslate this document into languages other than Englishfor nonprofit purposes. However, the ICMJE does nothave the resources to translate, back-translate, or approvereprinted or translated versions of the document. Thus,any translations should prominently include the followingstatement: ‘This is a (reprint /(insert language name) lan-guage translation) of the ICMJE Uniform Requirements forManuscripts Submitted to Biomedical Journals. (insertname of organization) prepared this translation with sup-port from (insert name of funding source, if any). TheICMJE has neither endorsed nor approved the contents ofthis reprint/translation. The ICMJE periodically updatesthe Uniform Requirements, so this reprint/translation pre-pared on (insert date) may not accurately represent the cur-rent official version at www.ICMJE.org. The official ver-sion of the Uniform Requirements for ManuscriptsSubmitted to Biomedical Journals is located at www.ICMJE.org.”

We do not require individuals or organizations thatreprint or translate the Uniform Requirements for Manu-scripts Submitted to Biomedical Journals to obtain formal,written permission from the ICMJE. However, the ICMJErequests that such individuals or organizations provide theICMJE secretariat with the citation for that reprint or


16 www.icmje.org

translation so that the ICMJE can keep a record of suchversions of the document.

IX. INQUIRIES

Before sending an inquiry, please consult FrequentlyAsked Questions at www.icmje.org, as this section of theWeb site provides answers to the most commonly askedquestions.

Inquiries about the Uniform Requirements should besent to the ICMJE Secretariat office by using the “ContactICMJE” link on the home page of www.icmje.org. Pleasedo not direct inquiries about individual studies, individualjournal styles, or individual journal policies to the ICMJEsecretariat office. The ICMJE does not archive individualjournal contact information. Manuscripts intended forsubmission to a journal must be sent directly to the jour-nal, not to the ICMJE.


www.icmje.org 17

Documents

Writing and Editing for Biomedical Publication