Reusing a patient cohort across two papers is legitimate. Concealing it is not. Disclose overlapping patient cohorts in your cover letter: name the related paper by DOI, state how many patients are shared, and state the distinct question each paper answers. Editors then judge whether the second paper stands alone.
Trouble starts with the discovery, not with the overlap. Forty shared patients out of ninety, disclosed at submission, are a decision for the editor. The same forty, found by a reviewer after acceptance, are an integrity question.
What counts as an overlapping patient cohort
An overlapping patient cohort is a set of participants, records, specimens or images appearing in more than one manuscript, whether published, in press, under review or posted as a preprint. Overlap is a fact about your data, not a verdict on your conduct. The ICMJE Recommendations call it overlapping publication and ask you to name any related manuscript at submission. What decides the verdict is whether the second paper asks a question the first did not answer.
| Situation | Status | What must be disclosed |
|---|---|---|
| A different outcome in a cohort you have already reported | Acceptable in principle when disclosed | The earlier paper, the number of shared patients, the new outcome |
| Extended follow-up of a cohort reported at an earlier time point | Acceptable in principle when disclosed | The earlier report, how many patients carried forward, the new follow-up window |
| A subgroup or secondary analysis of a parent study | Acceptable in principle when disclosed | The parent study, its registration number where it is registered, whether the analysis was pre-specified or post hoc |
| The same outcome in the same patients, written up again | Redundant publication | Not submissible as new work |
| One study divided so that neither half answers a whole question | Salami slicing | Combine the slices into a single paper |
The last two rows are the failure modes set out in duplicate and redundant publication, and salami slicing. The rows above are ordinary practice; disclosure keeps a submission there.
How editors detect overlap you did not declare
Editors detect overlapping patient cohorts through routine checks, and through specialist reviewers who read the journals you publish in.
- Similarity screening. A cohort description carried over near-verbatim makes your Methods paragraph match your own earlier article. How similarity screening works explains why a high self-match is a signal to investigate, not a plagiarism finding.
- Arithmetic in the Methods. The same recruitment window, centre and inclusion criteria, and a sample size that is a neat subset of a published series.
- Baseline tables. Identical age, sex and comorbidity distributions are hard to explain otherwise.
- Your own public record. An ORCID profile and a Crossref metadata search assemble your prior output in seconds.
- Your reference list. Citing the earlier paper without saying the patients are the same tells an editor where to look.
Detection after publication is the expensive version: COPE publishes flowcharts for suspected redundant publication, and the remedies run from a correction through an expression of concern to retraction.
The stand-alone test: can the second paper survive on its own?
Establish first that the second paper deserves to exist separately. Answer these six questions before you draft the disclosure; the answers to the first and the last go straight into the cover letter.
- Can you state a question the second paper answers and the first did not, in one sentence, without the word "also"?
- Does it have its own objective, analysis plan and conclusion, rather than a leftover result from the first?
- Would a reader of the first paper learn something they could not have inferred?
- Does the second abstract make sense to someone who has never seen the first paper?
- Would the two papers be a better single paper? If yes, combine them.
- Can you describe the overlap in numbers a stranger could verify from both papers' published Methods?
A no to question one, two or three means the paper is a slice, not a study. Question six separates a disclosure from a gesture: "there is some overlap with our previous work" gives an editor nothing to act on.
Cover-letter wording that discloses overlapping patient cohorts
A usable disclosure has four parts, each answering a question the editor would otherwise ask.
| Part | What to state | Worked example |
|---|---|---|
| The related paper | Full citation and DOI, plus status if not yet published | J Example Surg 2025;12:44-51, doi:10.[prefix]/[suffix] |
| The shared cohort | The source population, and the period over which it was assembled | A consecutive single-centre series of adults operated on between March 2021 and August 2023 |
| The extent of overlap | Numbers both ways: how many are shared, out of how many in each paper | 62 of the 118 patients here also appear in the 2025 paper, which reported 74 patients |
| The distinct question | The question and primary outcome of each paper, in parallel | The 2025 paper reported 30-day morbidity; this manuscript reports three-year recurrence |
Give the disclosure its own labelled paragraph. Copy and adapt the wording below; the rest of the cover letter follows the usual structure.
Dear Editor,
We submit "[Manuscript title]" for consideration as [an article type].
Disclosure of overlapping data. The patients in this manuscript overlap with work our group has already published: [Authors]. [Title]. [Journal] [Year];[volume]:[pages]. doi:[DOI]. That paper reported [N1] patients [treated / recruited / sampled] at [setting] between [start date] and [end date]. This manuscript analyses [N2] patients from the same consecutive series, of whom [N shared] appear in the earlier paper.
The two papers answer different questions. The published paper asked [question 1] and reported [primary outcome 1]. This manuscript asks [question 2] and reports [primary outcome 2], which was neither analysed nor reported in the earlier work. No table, figure or outcome from that paper is reproduced here. The cohort description in our Methods is necessarily similar and is cited to it at reference [n].
Other work from this cohort. [Delete if none.] A further manuscript from the same series, reporting [question 3], is under consideration at [journal]; we will inform you of the outcome. A preliminary version was presented as an abstract at [meeting, year] and not published in full.
The manuscript is not under consideration elsewhere, and all named authors have approved it.
Yours sincerely,
[Corresponding author, on behalf of all authors]
Where the disclosure also belongs inside the manuscript
The cover letter is not published, and because it names you it cannot go to reviewers under double-blind review. The overlap therefore has to appear in the manuscript too, written so it does not identify you.
- Methods, first paragraph. One sentence is enough: "Patients were drawn from a consecutive series reported in an earlier study [12]; this analysis reports outcomes not examined in that report." Write it in the third person, with no "we" and no "our previous", so the blinding holds.
- The participants item of your reporting checklist. The reporting checklist for your design, listed in the EQUATOR Network library — STROBE for observational work — asks for the eligibility criteria and the sources and methods of selection of participants. A previously reported cohort belongs there.
- Limitations. Cite the earlier paper, and say that shared patients make the two sets of results non-independent, which matters to anyone pooling them in a meta-analysis.
- Data availability statement. If the dataset is the one deposited for the earlier paper, point to that record rather than a second.
- Tables and figures. Do not reproduce them. If a baseline table must reappear, cite the source and check the licence: a CC BY article may be reused with attribution, a restrictively licensed one may not.
What we ask for at Directive Publications, and the page we do not have
Our author guidelines already ask that your cover letter disclose any related or prior work and preprints. An overlapping patient cohort is related work, so the disclosure is a requirement rather than a courtesy, and the cover letter is where it is recorded. Similarity screening runs on submissions, review is double-blind, at least two independent expert reviewers are sought for research manuscripts, and the handling editor decides.
Being plain about what is and is not published: our research integrity policy states that redundant or duplicate publication, undisclosed overlap with the authors' own prior work, and salami-slicing a single study into minimal publishable units are not acceptable; our publication ethics policy covers redundant, duplicate and concurrent submission; and our corrections policy sets out what happens to an article after publication. What we do not publish is a page devoted to overlapping cohorts specifically, or any threshold at which shared patients make a second paper redundant — that judgement stays with the handling editor. If your case is unusual, put it to the editorial office in writing before you submit and keep any reply.
The boundary is set out in duplicate and redundant publication; the letter itself in how to write a cover letter to the editor; what a screening report shows an editor in how similarity screening works; and who signs the letter in the corresponding author's responsibilities.