1. The editorial purpose of post-publication action
Publication does not end editorial responsibility. When credible information calls a published article into question, the journal must assess whether readers need clarification, a correction, an interim notice or withdrawal of confidence in the work. The objective is an accurate and intelligible scholarly record, not punishment of authors or protection of the journal’s reputation.
Post-publication action should respond to the reliability and interpretation of the published material. The editor must distinguish an error that leaves the principal conclusions intact from a defect that makes the article’s findings unreliable. A misconduct finding may be relevant, but a journal does not need to establish intent before addressing an unreliable record.
Apply the publisher’s mandatory corrections-and-retractions policy alongside journal-specific procedures. Assign a responsible case handler, preserve independence and record the basis for each decision.
2. Recognize and triage post-publication concerns
Concerns may arrive from authors, readers, reviewers, institutions, publishers, journalists or anonymous correspondents. They can involve numerical inconsistencies, image problems, unreported overlap, authorship, ethics approval, consent, undeclared interests or errors introduced during production. Classify the observable issue before assigning it a label.
At intake, locate the exact article, version, table, figure, dataset or statement at issue. Assess whether the report includes verifiable details and whether the potential impact is confined to a minor passage or extends to the central evidence. Anonymous reports should be assessed on their substance rather than dismissed because their source is unknown.
Urgent issues—such as disclosure of identifiable participant information or serious risks arising from unreliable clinical claims—may require immediate protective steps. Urgency changes the timetable; it does not remove the need for fairness and accurate communication.
3. Preserve evidence and establish case governance
Secure the version of record, submitted and accepted files, production proofs, relevant figures, underlying correspondence and any supplied evidence. Preserve timestamps and version histories where available. Internal notes should separate observations, allegations, author explanations, expert opinions and established facts.
Identify who will handle the case and who has authority to approve notices. An editor implicated in the concern or closely connected to affected authors should disclose the relationship and follow the journal’s recusal procedure. Publisher-level oversight may be necessary for complex or cross-journal cases.
Restrict access to case materials, especially unpublished files, confidential reviews, personal information and institutional correspondence. Do not circulate allegations through informal mailing lists or upload confidential evidence to unapproved external tools.
4. Assess materiality and the reliability of the published conclusions
The central editorial question is how the identified issue affects what readers can reasonably rely on. A typographical error in a nonessential value, a mislabeled panel, a mistaken statistical calculation and unavailable primary data can have very different implications. Assess the issue in the context of the article’s methods, analyses and claims.
Ask whether the error is localized and demonstrably correctable, whether the corrected information changes interpretation, and whether other results independently support the central conclusions. A narrow correction should not be used when the published evidence as a whole cannot sustain the article’s main claims.
Where the editor lacks necessary expertise, obtain focused independent statistical, methodological, image or subject-matter advice under appropriate confidentiality arrangements. Consultants should evaluate evidence and reliability, not speculate about intent.
5. Seek explanations and supporting materials fairly
Inform the appropriate authors of the specific concerns and request a focused response, relevant source material or documentation. The request should explain what needs clarification and provide a reasonable opportunity to respond. A request for evidence is not an accusation of misconduct.
Where the concern affects only one part of the work, request material proportionate to that question. If original data or image files are needed, specify the purpose and use an approved secure route. Avoid requesting identifiable participant data when a de-identified or institutionally verified alternative would suffice.
Evaluate explanations against the available record. Incomplete documentation may leave reliability unresolved even when misconduct is not established. Record follow-up attempts and distinguish nonresponse from a finding about the underlying allegation.
6. Choose between correction, expression of concern and retraction
A correction is generally used when a specific error can be repaired and the corrected article remains sufficiently reliable. An expression of concern can alert readers to a material unresolved issue while a credible inquiry remains incomplete or inconclusive. Retraction is appropriate when the published findings or conclusions cannot be relied upon, or when another serious publication problem meets the applicable policy threshold.
These notices serve different purposes and should not be selected to minimize reputational discomfort. A correction cannot rescue conclusions unsupported by the corrected evidence; an expression of concern should not become a substitute for completing a reasonably achievable assessment; and a retraction should not be framed as a determination of author intent without an appropriate evidentiary basis.
Some situations warrant no public notice after a documented assessment, while others require a coordinated combination of actions affecting linked publications or datasets. Explain why the selected action addresses the actual risk to readers.
7. Issue substantive corrections that are precise and traceable
A correction should identify the original article, describe the error, state the corrected information and explain any effect on interpretation. Where tables, figures, supplementary files or analyses change, readers should be able to locate the corrected material without reconstructing it from vague prose.
Decide whether the correction can be expressed in a linked notice, requires an updated article file, or requires both. Maintain a transparent version history and make clear which parts of the record have changed. Silent replacement of substantive published content undermines traceability.
Check that revised text, numbers, figures, supplementary materials and conclusions are internally consistent. A correction should itself undergo appropriate editorial and technical verification before publication.
8. Use expressions of concern as carefully scoped interim notices
An expression of concern can be appropriate when credible evidence raises a material question about reliability but the available information does not yet justify a final corrective action. It informs readers of uncertainty rather than declaring that misconduct occurred.
The notice should identify the affected article, describe the concern to the extent supported and explain the status of any ongoing assessment that can appropriately be disclosed. Avoid presenting contested allegations as proven facts or revealing confidential investigation details.
Assign responsibility for periodic review of the notice. When new evidence arrives, determine whether to maintain, update or replace it with a correction, retraction or another documented outcome. Interim notices should not be abandoned indefinitely.
9. Make and communicate a retraction decision
Retraction addresses a publication that should not continue to be relied upon as part of the scholarly evidence base under the applicable policy. Grounds can include unreliable results arising from major error or other causes, redundant publication that materially misrepresents the record, or serious ethical or publication problems warranting withdrawal of the article’s claims.
Assess the article as a whole and consider whether a narrower correction can genuinely restore reliability. Do not require proof of deliberate misconduct when the evidence establishes that the central findings are unreliable. Conversely, avoid retracting an otherwise reliable article solely because of an unrelated minor error.
Document the evidence, author responses, specialist input and rationale. The retraction notice should explain the reason accurately and avoid unsupported statements about personal culpability.
10. Write transparent notices without unnecessary accusations
Every public notice should be intelligible as a standalone document. Identify the affected article using stable bibliographic information and a persistent link, state the action taken and describe the issue with enough specificity to inform readers. The notice should make clear whether the conclusions remain reliable, are under question or can no longer be relied upon.
Separate established findings from unresolved allegations and identify the responsible party for the notice according to journal practice. If some authors agree and others disagree, document the response accurately without implying unanimity.
Use restrained language and avoid unnecessary personal details. Public notices correct the literature; they are not venues for adjudicating employment disputes or publishing confidential correspondence.
11. Preserve the article, metadata and links across platforms
Retracted articles should generally remain accessible as part of the historical scholarly record, clearly and prominently marked as retracted, with a direct link to the retraction notice. Removing the article entirely can obscure what was published and complicate the interpretation of subsequent citations, subject to exceptional legal or privacy circumstances.
Ensure that article landing pages, PDF or HTML versions, correction or retraction notices, issue listings and relevant metadata consistently communicate the current status. Work with production and indexing partners to propagate updates where the journal has the necessary arrangements.
Maintain durable links between the original article and each notice. When several updates occur, readers should be able to reconstruct the sequence without guessing which version is authoritative.
12. Coordinate with institutions and other journals
Institutions may be able to examine primary records, ethics oversight or matters beyond the journal’s authority. When referral is justified, use an authorized channel and provide a factual description of the concern and relevant evidence without presenting allegations as established misconduct.
An institutional investigation and a journal’s publication decision are related but distinct. The journal may need to protect readers before an external process concludes, or may reach a publication-related decision despite an inconclusive institutional response. Record the basis and limitations of the journal’s own assessment.
When multiple articles or journals share affected data, images or a compromised review process, coordinate through appropriate publisher and journal channels. Share only information necessary for the specific concern and respect confidentiality.
13. Handle author-initiated corrections and disagreements
Authors should have a clear route to report errors discovered after publication. A voluntary disclosure should prompt the same materiality assessment as a concern from any other source. Timely cooperation can help the journal clarify the record but does not predetermine the appropriate public action.
Where authors disagree about the facts or the proposed notice, provide a fair opportunity for relevant responses while maintaining the journal’s authority over the published record. Document the differing positions and avoid assuming that the corresponding author can speak conclusively for all coauthors in a contested case.
A journal may need to act when an article is demonstrably unreliable even without unanimous author agreement. The notice should describe the evidence and action rather than using the dispute itself as the sole justification.
14. Manage ethical, privacy and consent-related corrections
Post-publication concerns involving human participants, consent, identifiability or sensitive information require careful assessment of both reliability and potential harm. A published image or dataset may create a privacy risk even when the article’s scientific conclusions remain otherwise sound.
Use the publisher’s authorized privacy, ethics and incident-response procedures when immediate safeguards are needed. Restrict unnecessary circulation of sensitive material and avoid contacting research participants directly as an improvised editorial investigation.
The appropriate response may involve correcting or restricting specific material, adding a notice, or considering broader publication action according to the severity and the governing policy. Exceptional removal of content should be documented and distinguished from ordinary retraction.
15. Address image, data and statistical concerns after publication
An image irregularity may reflect an assembly error, reuse requiring explanation or a more consequential problem affecting the underlying evidence. Identify the specific panels, comparisons and claims at issue before deciding whether a corrected figure would resolve the concern.
For statistical or data discrepancies, ask which reported results change after correction and whether the article’s principal interpretation remains supported. Independent technical advice may be necessary to assess reproducibility or the implications of missing source material.
If replacement figures or analyses are submitted, verify that the revised material is appropriately supported and does not introduce inconsistencies elsewhere in the article. A cosmetic change should not be accepted as a substitute for resolving a material reliability concern.
16. Treat overlap, authorship and competing-interest issues proportionately
Duplicate or overlapping publication requires assessment of the relationship between the affected works, prior disclosures, reused data and the significance of the overlap. Permitted preprints and transparently linked secondary analyses should not be conflated with undisclosed redundant publication.
Authorship disputes may require institutional determination of contributions or consent. Editors should avoid adjudicating workplace disputes beyond the journal’s authority while ensuring that the published author record is accurate to the extent verifiable.
An omitted competing-interest declaration may require an update or correction, but the editor should separately assess whether it affected the interpretation, editorial process or reliability of the article. Different publication problems should not be collapsed into a single automatic remedy.
17. Manage timing, communications and procedural fairness
Act promptly enough to protect readers, but allow a reasonable opportunity for authors and other relevant parties to respond. The appropriate timeline depends on urgency, complexity, specialist consultation and whether an external inquiry is underway.
Keep communications factual and coordinated through the assigned case handler. Explain the procedural status where confidentiality permits, without promising an outcome or revealing confidential sources and internal deliberations unnecessarily.
If a public notice is delayed by unresolved evidence, assess whether an interim statement is needed. A lack of certainty should be communicated as uncertainty, not silently converted into inaction when readers face a material risk.
18. Review corrections and retractions for accessibility and discoverability
Public notices should be easy to find from the original article and understandable without specialized knowledge of the journal’s internal process. Clear titles, direct links and consistent presentation help readers recognize whether they are viewing corrected, questioned or retracted material.
Check that the notice appears in the journal’s article pages and relevant formats and that internal search and navigation do not conceal the updated status. Production and metadata teams should verify links, identifiers and any applicable notifications to indexing services.
Where the journal hosts multiple versions, identify the version of record and maintain a clear change history. Accessibility and technical accuracy are part of the corrective action, not optional finishing steps.
19. Close the case and improve post-publication oversight
At closure, preserve the original concern, evidence considered, author responses, specialist advice, conflict-management steps, decisions, notices and relevant communications in the approved system. Record any unresolved limitations and which external referrals, if any, remain open.
Verify that the public record reflects the final action across the formats and services within the journal’s control. A signed decision without correctly linked and visible notices does not fully protect readers.
Review recurring patterns to improve submission checks, data and image guidance, reviewer training, production verification and post-publication reporting routes. Use documented lessons rather than unsupported assumptions about individual motives.
20. A repeatable post-publication decision framework
First define the observable concern and secure the published and underlying records. Assess materiality, urgency, conflicts and the limits of journal authority. Seek focused explanations and expert advice when needed, distinguishing verified facts from allegations.
Next determine whether the article remains reliable, can be made reliable through a transparent correction, requires an interim expression of concern or meets the journal’s grounds for retraction. Select the action that accurately informs readers rather than the one that appears administratively easiest.
Finally write a precise notice, preserve the article-to-notice links, update the formats and metadata the journal controls, and document the outcome. The standard is a fair, proportionate and discoverable correction of the scholarly record.