Editorial Policies

Editorial decisions are based on scope, scientific quality, methodological rigor, ethical compliance, reporting quality, and relevance to the journal.

The Editor-in-Chief is accountable for the journal's editorial direction and for the decisions taken under it. Handling editors work within these policies, and the publisher does not intervene in the assessment of individual manuscripts.

Model

Single-blind by default; double-blind review is available on request.

Screening

All submissions undergo initial editorial screening. Manuscripts that meet the journal's scope and minimum requirements proceed to independent peer review.

Reviewers

Manuscripts that proceed to external peer review are normally evaluated by at least two independent subject-matter experts.

A request for double-blind review is made at submission. When it is granted, editors ensure that author-identifying information is not shared with reviewers, and reviewers are asked not to attempt to identify the authors. The preparation this requires of authors is set out in the Instructions for Authors.

Reviewers advise; the handling editor decides. Where reviews conflict, the editor weighs them against the evidence in the manuscript rather than counting recommendations, and may seek an additional opinion. The operational detail for each side of this is on the Editor Guidelines and Reviewer Guidelines pages.

Authorship is limited to those who made a substantial contribution to the conception, design, data acquisition, analysis or interpretation of the work; who drafted or critically revised it; who approved the version submitted; and who accept accountability for its accuracy and integrity. Everyone meeting these conditions is listed, and no one who does not meet them is.

Contributions that do not meet the criteria — funding acquisition alone, general supervision, technical assistance, language editing, data collection without interpretation — are recognized in an acknowledgments section with the named person's permission.

Any change to the author list after submission requires written agreement from every author, including the person being added or removed, together with the reason for the change. The corresponding author is responsible for obtaining it.

Generative AI tools are not authors. They cannot take accountability for the work, and any use of them is disclosed as described in section 07.

Originality and prior publication

Submitted work must be original and must not be under consideration elsewhere. Deposit of a preprint is not prior publication and does not preclude submission; the preprint and its DOI should be declared. Substantial overlap with the authors' own earlier work is declared at submission, with the earlier work cited, so the editor can judge what is new.

Similarity screening

Submissions are screened with Crossref Similarity Check, powered by iThenticate. A similarity report is evidence for an editor to interpret, not an automatic threshold: matched text is assessed for what it is — a methods sentence reused from the authors' own protocol is not the same finding as an undisclosed reproduction of another group's discussion.

Data and image integrity

Fabrication and falsification of data are grounds for rejection and, after publication, for retraction. Figures must represent the underlying data: adjustments to brightness, contrast or color are applied to the whole image and must not obscure, move, introduce or remove any feature. Where an image is assembled from separate acquisitions, the composition is stated in the legend. Editors may request original, unprocessed files at any stage.

Concerns raised about a published article are investigated following COPE guidance, with the authors and, where appropriate, their institution given the opportunity to respond before any action is taken.

Studies involving patient data require documented approval from an institutional ethics committee or IRB, named in the manuscript with its approval reference. Consent is documented for prospective studies; where a retrospective analysis proceeded under a waiver, the waiver and the body that granted it are stated.

Methods for de-identifying patient data must be described clearly enough for a reader to judge the residual risk of re-identification. Studies conducted in the United States demonstrate HIPAA-compliant data handling; studies conducted in the European Union comply with the General Data Protection Regulation. Individual patients must not be identifiable in text, tables, figures or supplementary files, including through free-text clinical notes, rare-diagnosis combinations, dates or geographic detail. Where an identifiable image or case description is essential, written consent for publication is obtained and stated.

Clinical-trial registration

Interventional clinical trials must be registered in an accepted public trial registry at or before the first participant provides consent for enrollment. The registry name, registration number, and registration date must be reported in the manuscript.

Every empirical research manuscript must include a data-availability statement.

The statement says where the data are, on what terms, and what a reader must do to obtain them. Clinical data that cannot be shared openly are described honestly: the governance that restricts them, the access process, and the contact point for a request. "Available on request" without a named contact and a described process is not a data-availability statement.

Custom algorithms and models should be deposited in a public repository, documented well enough to replicate the computational methods, with the version or commit used for the reported results identified. Where code cannot be released, the manuscript states why and describes the method in enough detail for independent reimplementation.

Data provenance is part of the method: the source system, the extraction period, the inclusion and exclusion criteria, the handling of missing values, and any transformation applied before analysis.

This journal publishes work intended to influence clinical decisions. A claim about clinical benefit is therefore held to the evidence that supports it, and not to the performance of the model that produced it.

Studies developing clinical decision support or AI-driven diagnostics address algorithmic bias and fairness across demographic groups, the intended user and decision context, interpretability appropriate to that user, and the human oversight assumed in deployment. Validation, comparator selection and data leakage are assessed at review; the reporting expected of each study type is set out in the Instructions for Authors.

AI-assisted writing and analysis

Authors disclose any use of generative AI tools in preparing the manuscript or in conducting the analysis, naming the tool and describing what it was used for. Authors remain fully responsible for the content, including anything an AI tool produced. Reviewers are subject to a separate rule, set out on the Reviewer Guidelines page.

Authors disclose financial relationships with health IT companies, electronic health record vendors, device manufacturers or AI companies whose products are discussed or evaluated, including employment, consultancy, equity, patents, paid advisory roles and speaking fees. Non-financial interests that a reader could reasonably regard as competing are disclosed on the same terms. All funding sources are named, with grant numbers, together with a statement of what role the funder had in design, analysis, interpretation or the decision to publish.

Editors and reviewers declare competing interests and recuse themselves where one exists. Where a member of the editorial team is an author, the manuscript is handled by an editor who is not, and the author has no access to its review or decision.

The published record is permanent. An article is not removed, replaced or silently edited after publication; where something is wrong, the correction becomes part of the record and stays linked to the article it concerns.

Notices follow COPE guidance, state what is being corrected and on whose initiative, and are published open access with their own identifier. A retracted article remains available and remains listed in the Archive, marked so that it cannot be mistaken for ordinary research.

An author who believes a decision rests on a misunderstanding of the work may appeal in writing to the editorial office, addressing the specific points in the reviews and reports. Appeals are considered by an editor who was not responsible for the original decision, who may seek further review. A single appeal is considered per manuscript, and the outcome of that appeal is final.

Complaints about editorial process, conduct or publication ethics are made to [email protected] and are investigated by someone not involved in the matter complained of. Where a complaint concerns the Editor-in-Chief, the publisher appoints an independent reviewer.

An article processing charge applies after acceptance. It has no bearing on the editorial decision, and the published rates and the terms attached to them are on the Article Processing Charges page.

Authors should contact the editorial office promptly and wait for written confirmation before submitting the work elsewhere. See the complete withdrawal policy.

Articles are published open access under the Creative Commons Attribution 4.0 International license (CC BY 4.0 (opens in a new tab)). Authors retain copyright in their work. Anyone may read, download, copy, distribute, adapt and build on a published article, in any medium, provided the original work is properly cited.

Material reused from another source — a figure, a table, an instrument, an extended quotation — requires written permission compatible with that license, obtained by the authors before submission and acknowledged in the caption.

How the published record is identified, discoverable and preserved is described on the Indexing and Discoverability page.