Author guidelines

Submit in the format you already write, in minutes. You upload the manuscript, the system reads it and pre-fills the forms, and you verify rather than re-type. What we ask for is a reproducibility package, as complete as your data allow. It is what this journal is built on. Where privacy limits what can be shared, your paper still earns a badge, marked for what could and could not be verified. Everything you upload must be de-identified before it leaves your machine: the requirement is set out below, and it admits no exceptions.

Formats

• Quarto, R Markdown or Jupyter (preferred). One executable source is the strongest reproducibility substrate and earns the full verification treatment.
• LaTeX (accepted). Code and data are supplied as a separate bundle.
• Word (accepted). Code and data go in a separate bundle, and the article earns a reproducibility badge reflecting what we could verify.

There is no forced template. Whatever tool you write in, we turn it into the published HTML and PDF. You focus on the science.

How to submit

Sign in at submit.fairpressjournals.com with your email address. A 6-digit code is emailed to you, no password required. Upload the manuscript as a PDF — or, if you wrote it in Word, the .docx itself. The form then asks for the manuscript source: the markup file the published article is typeset from (.tex, .qmd, .rmd, .ipynb, or a zip of a LaTeX project). A Word .docx manuscript is its own source, so Word authors upload no separate source file. Code and data go in a bundle (zip) or a linked repository, not in the manuscript file. The system scans the upload and pre-fills the title, abstract, keywords, authors and more as clearly marked suggestions. You confirm each one and add anything missing.

What submission requires

• Article type: new method, applied, simulation, replication, null or negative finding, theory, or software.
• Title, abstract (up to 400 words), and 4–6 keywords.
• Authors: every author with name and affiliation, and exactly one corresponding author with a reachable email address. ORCID iDs are optional at submission (validated if given) and required before acceptance.
• The manuscript source, separately from the manuscript file: the markup the article is typeset from (.tex, .qmd, .rmd, .ipynb, or a zip of a LaTeX project — a .docx manuscript counts as both). Applied, new-method and simulation papers must supply it; replication, null-finding, theory and software papers may instead add a short note explaining why none exists.
• The reproducibility package. For new-method, applied and simulation papers, submission requires the code's location (bundle or repository) and the one entry-point command that regenerates the results, for example Rscript run_all.R or make. The environment specification, data details and determinism (seeds) are collected too and confirmed before acceptance. Replication, null-finding, theory and software papers provide the artifacts that exist for them, or a documented exemption with the protocol followed. See data & code availability.
• Replications additionally name the target work (DOI) and the replication protocol.
• Disclosures: data and code availability, AI use, funding, competing interests, and ethics approval where applicable. You choose from set options, and only the disclosures that apply to your paper appear. Each needs an answer, and a declared "none" is a valid one.
• Two attestations: that the reference list is real and checked, and that the entered metadata matches the manuscript.

The items below are specific to health research, and the submission form has no field for most of them. Put them in the manuscript or an appendix to it: the manuscript document is what the automated assessments read. The bundle and the repository are available to the editor, who consults them where they bear on the decision.

Which article type to choose

The type list is shared across all FAIR Press Journals titles, so health genres map onto it rather than appearing by name. A trial, a cohort or case-control study, a diagnostic-accuracy study or a systematic review is normally applied; a reanalysis of a named prior study is replication; a paper whose contribution is a new statistical or modelling method is new method; a methods paper built on simulated data is simulation. The type decides which reproducibility items are required at submission, so if two of them fit, pick the one that describes your contribution and say in the cover text why.

Submit de-identified material only

Nothing you upload may identify a patient, participant or any other individual. This covers the manuscript and everything shipped with it: tables, figures, appendices, code, log files, and any dataset or surrogate in the bundle or the linked repository.

• No direct identifiers. Names and initials, addresses, telephone numbers and email addresses, medical-record, insurance or social-security numbers, full dates of birth and dates of death, admission and discharge dates precise enough to single someone out, ages above 89, device, account and IP identifiers, audio or video recordings, biometric and genotype data, facial photographs or scans and any other image from which a person could be recognised — a black bar over the eyes or a blur is not de-identification — and free-text notes that describe a person recognisably.
• Watch the indirect identifiers too. A rare diagnosis in a named clinic, a table cell containing one or two people, geography finer than the population supports, an unusual combination of age, occupation and date: individually innocuous, jointly identifying. Aggregate or suppress small cells before upload rather than after review.
• Check the artifacts, not just the paper. Clear notebook outputs before you commit or zip: a saved .ipynb carries the rows it printed. Check the repository history too — deleting a file does not remove it from the history a reader can clone — and look through logs, cached intermediate files and figure source data.
• Quotations and case-level detail in a mixed-methods paper are paraphrased or redacted to the point where the speaker cannot be recognised, and are covered by the consent under which they were collected.
• If your analysis cannot survive de-identification, do not upload the data. Take the restricted-data route on data & code availability, or write to the editors before you submit and we will tell you what we can review.

Be exact about this, because we are exact about what happens either way. Your manuscript is assessed by a disclosed, AI-assisted engine running on a commercial model API under a data processing agreement with EU standard contractual clauses: it is not used to train models and is deleted under the contractual retention limit. Every queued manuscript takes that path before a human reads it — and so does the upload scan described above, which sends the first pages of your file to the same contracted API the moment you upload it, before you press Submit. A file uploaded with identifiable data in it has therefore already been processed by the time anyone notices. If identifiable individual-level data is found in a submission, at any stage and by anyone, the editor stops the assessment, has the material purged from our storage, requests its deletion under the contract's retention rule, and contacts the corresponding author. That is an incident response, not a safety net: only you can keep the data out. The separate rule for the raw datasets handled by the imprint's optional safe-companion-dataset service — that they never leave a controlled environment and never touch a cloud AI model — is described on data & code availability and is about those datasets, not about manuscripts.

Interventional studies: prospective registration

Trials and other interventional studies are considered only if they were registered prospectively. As a condition of consideration, the study must be registered in a WHO ICTRP primary registry or ClinicalTrials.gov at or before the time of first patient consent for enrolment, and the manuscript must state the registration identifier, the registry and the registration date, together with the date of first patient consent or enrolment, so the two can be compared. There is no exception to the timing requirement: a study registered after enrolment began cannot be considered here, however good it is.

Review checks, from the manuscript alone, that a registration is stated, that it is prospective against the paper's own dates, and that the outcome reported as primary is the one the paper presents as registered. The registry entry itself is then checked by the editors against the registry before a decision is signed. It is not checked automatically, and we do not claim it is. Where the registered and reported primary outcomes differ, explain the change and when it was made: an explained deviation is a normal part of research, an unexplained one is not. Study protocols are not an article type here; SPIRIT 2025 is named only as the reference for what a registered protocol should contain.

Reporting guidelines: ship the completed checklist

Follow the EQUATOR-matched reporting guideline for your design, and include the completed checklist as an appendix to the manuscript, with the page or section where each item is reported. A copy in the bundle is welcome, but the appendix is the copy the assessments read:

• CONSORT 2025 — randomised trials.
• STROBE — observational studies, with RECORD for studies using routinely collected health data.
• PRISMA 2020 — systematic reviews and meta-analyses.
• TRIPOD+AI — development and validation of clinical prediction models.
• STARD — diagnostic-accuracy studies.
• CHEERS 2022 — economic evaluations.

Use the current versions named here. Where no listed guideline fits your design, say which one comes closest and where your study departs from it. A checklist makes a paper checkable, which is the same thing this journal asks of code and data. Registering a review protocol (PROSPERO or similar) and stating the registration is encouraged and is not a condition of consideration — prospective registration is required for interventional studies only.

Data that cannot be shared, and the badge you get

Patient records, registers, linked administrative data, data held under a use agreement: a real constraint, and not a reason to stay away. Reproducibility here is a default, not a barrier. What we ask is that the restriction is declared — what the data are, why they cannot be released, and the route by which another researcher could apply for access — and that you ship what can be shared: the analysis code, the derived tables the results rest on, and, wherever one can be built, an anonymised or synthetic surrogate the analysis can be rerun on.

The article then carries the badge the check actually earned. ✓ Reproduced means the deposited code, over the deposited data, regenerated the results as printed; ◑ Partially reproduced means some of them did; Privacy-limited means the underlying records could not be released, and it is a distinct, honest state rather than a lesser one. An article carries one badge, not a stack: on a restricted-data route the badge is Privacy-limited whatever the surrogate rerun produced, and the rerun is reported in words in the article's reproducibility panel, which always says what was verified and what was not. The routes, and the optional service that can build a surrogate for you, are on data & code availability.

Mixed-methods submissions

Mixed-methods work with a quantitative core is welcome. Its qualitative strand is read by the editor for transparency of method and material — how participants were selected, how the material was collected and coded, whether the coding frame and (de-identified) extracts are available. The assessment written for this journal's designs, which judges design and inference, is instructed to assess the quantitative core only and never to demand a power calculation or an effect size of the qualitative strand; the other assessments read the manuscript as a whole, so write that strand for a quantitative reader. Purely qualitative studies are out of scope in this first version, for reasons stated plainly on aims & scope: the review roster this journal runs is quantitative, and we would rather build the support first than review such work badly.

What happens next

Incomplete submissions come back with the specific missing items named, not a generic rejection. Complete submissions enter review, with disclosed AI-assisted assessment, a reviewer panel, reproduction of your results, and a decision read and signed by a named editor. The full process is in the peer-review policy. Review is built to be fast. We publish our actual decision times on the journal page once enough decisions exist.

Fees

There is no submission fee. Publication is funded by an APC on acceptance, with need-blind waivers. Amounts and rules are on the fees & waivers page.

Stuck?

Write to contact@fairpressjournals.com. An editor reads it.