How to review a case report comes down to five checks in order: is it a case report, does it teach something, is the diagnosis established, is consent for publication stated, and is the reporting complete and consistent? Sort each problem as fatal or fixable, let the worst set your recommendation, and write each fixable one as a numbered request.
The general method — confidentiality, two readings, a structured report — is in our first-time peer reviewer's guide. Review models such as double-blind review are explained in our introduction to peer review. This guide covers what changes when the manuscript describes one patient.
Your reference standard is CARE, the reporting guideline for case reports, listed on the EQUATOR Network's CARE page. The 2017 CARE explanation and elaboration paper says it is designed to support critical appraisal by editors and peer reviewers, not only writing by authors. The items are covered from the author's side in our guide to writing a case report.
How to review a case report in five checks, in this order
A manuscript that fails one of the first three checks cannot be rescued by complete reporting, so do not audit every CARE item first.
- Is it a case report? Usually one patient, written up because something departs from what clinicians expect. CARE defines a case report as describing a problem in "one or several patients", so the patient count alone does not make a case series; our comparison of case reports, case series and cohort studies explains what does.
- Does it teach something? Write the lesson in one sentence. CARE asks the introduction to say why the case is unique (item 4) and the conclusion to give the take-away lesson (item 11d).
- Is the diagnosis established? See the next section. A lesson built on an uncertain diagnosis teaches the wrong thing.
- Is consent for publication stated? Look for the statement, and read the text and figures for details that would let someone who knows the patient recognise them, as set out in our guide to patient consent for case reports.
- Is the reporting complete and consistent? Work through the CARE items, then cross-check as described below.
Test the diagnosis and every alternative explanation
CARE item 8c asks for the diagnosis "including other diagnoses considered". For each condition that could produce the same picture, find the finding that excludes it. An alternative dismissed without a finding is still open.
Then ask what confirms the final diagnosis: histology, culture, a specific test, imaging, or a course only that diagnosis explains. If confirmation rests on response to treatment alone, the authors should say so as a limitation.
Last, separate coincidence from cause. Two events in one patient may simply share a timeline. CARE item 11c asks for the rationale for conclusions, "including assessment of possible causes", and its authors acknowledge that causality is a challenge for case reports. Check that the timing fits and that any mechanism is offered as a hypothesis; for a suspected drug reaction, use the causality checks in our guide to adverse drug reaction case reports. The conclusion may claim that something occurred, not how often, how risky or how effective it is.
Check that the timeline, text, abstract and figures agree
A case report gives you no dataset to check, so internal consistency is your main evidence of careful reporting. Use the timeline (CARE item 7) as the reference and test everything against it.
- Every date or interval appears in the narrative with the same value and order.
- No test precedes the finding that prompted it, and each change of treatment has a stated reason (item 9c).
- Values and units match across text, tables and figure legends.
- Every number in the abstract appears in the full text; the other abstract checks are in our guide to case report abstracts.
- Follow-up is long enough for the outcome claimed. CARE sets no minimum, so judge it against the condition and ask for the interval beside any claim of cure.
What not to demand of a case report
The Committee on Publication Ethics (COPE) says in its Ethical Guidelines for Peer Reviewers that extending the work beyond its scope is not the reviewer's job. Say which requests are essential to the claims and which would only strengthen the work.
| Do not ask for | Ask for instead |
|---|---|
| A test that was never indicated, or cannot now be done | Why it was not done (CARE item 8b covers diagnostic challenges such as access to testing), named as a limitation only if its absence leaves the diagnosis open |
| Statistics or a control group | Wording that claims no more than one patient can show |
| Re-exposing the patient to a suspected cause | Nothing; a patient is not re-exposed to strengthen a paper |
| A review of every similar case | The closest previous reports, and how this case differs |
| Ethics-committee approval as a condition of your recommendation | Nothing directly: whether one case needs committee approval depends on local rules and journal policy, so a missing or unclear ethics statement goes to the editor (see ethics approval and informed consent). Consent for publication is the requirement every case report must meet |
| Cuts to meet a word limit | Nothing; the journal's instructions set limits |
Fatal or fixable: match each problem to a recommendation
COPE asks that your recommendation match the comments in your report. Sort each problem below and let the worst row decide.
| What you find | Fatal or fixable | Recommendation |
|---|---|---|
| All five checks pass; only typing errors remain | Nothing substantive | Accept |
| Other missing CARE items, such as the patient perspective | Fixable | Minor revision |
| Identifying details the science does not need | Fixable | Minor revision, plus a line to the editor |
| A causal or "effective" claim one patient cannot support | Fixable | Minor, or major if the conclusion rests on it |
| Follow-up too short for the outcome claimed | Fixable | Minor or major revision |
| Alternatives not discussed, though data to address them exist | Fixable | Major revision |
| Timeline missing, or contradicting the text | Fixable if records allow | Major revision |
| A "first reported case" claim contradicted by an earlier report | Fixable | Major revision |
| No consent-for-publication statement | The editor's call | Flag to the editor; recommend on the rest |
| Suspected duplicate publication, image manipulation or fabrication | Not yours to judge | Confidential note; do not investigate |
| Diagnosis unconfirmed, no longer confirmable, and the lesson depends on it | Fatal | Reject |
| No teaching point that revision could supply | Fatal | Reject |
If you suspect the patient was published before, or an image was altered, tell the editor confidentially, as COPE and our Reviewer Guidelines ask, and leave the editor to judge duplicate publication, including whether an earlier abstract or preprint counts. If the case reveals who the authors are despite anonymisation, and that raises a conflict, tell the editor as soon as you realise it.
Model comments for the report, and a note to the editor
Before commenting on novelty, run one quick search in a free database. If an earlier report turns up, ask the authors to cite it and restate their claim as a dated statement about their own search, as our guide to "first reported case" claims explains.
The comments below are invented to illustrate the structure; they describe no real manuscript or patient.
Major 1. The diagnosis of [condition] rests on [finding]. The Discussion does not say how [alternative 1] and [alternative 2] were excluded. Please give the finding that excludes each, or state that one could not be excluded and moderate the conclusion.
Major 2. The timeline places [event] on [day], but the Case presentation describes it after [later event]. Please reconcile the two and check the figure legends against the corrected sequence.
Major 3. The title calls this the first reported case. A search of [database] on [date] found [reference], describing [similar case]. Please cite it, explain how this case differs, and restate the claim in terms of the search you ran.
Minor 1. The Discussion calls [treatment] effective. One patient cannot show effectiveness; please write that the patient [improved] after [treatment].
Minor 2. Please state the follow-up interval in the abstract, and avoid "cured" unless follow-up covers the period in which recurrence would be expected.
If a fatal problem leads you to recommend rejection, say why revision cannot remove it.
Major 1 (recommending rejection). The diagnosis of [condition] rests on [finding] alone. The [confirmatory test] was not done and cannot now be done, so [alternative] remains possible. The take-away lesson depends on that diagnosis, so I do not think revision could fix this.
COPE says confidential comments are not a place for denigration, so keep criticism of the science in the report.
Confidential comments to the editor. I recommend major revision. The case has a clear teaching point, and [test] supports the diagnosis; my concerns are fixable. Three points are for you. I found no statement that consent for publication was obtained. Figure [n] shows [feature], which could let someone who knows the patient recognise them. I did not assess the [histology or imaging], which is outside my expertise.
How case reports are reviewed at Directive Publications
Our Article Types page states that case reports receive external double-blind review at a depth appropriate to their scope. The handling editor decides and handles all contact with authors; reviewers advise.
Our Reviewer Guidelines suggest a report structure — Summary, Major issues, Minor issues, Confidential comments to the editor — and ask for a recommendation of accept, minor revision, major revision or reject. They ask you not to upload any part of a submission to generative AI tools, and to tell the editor if you used AI only to polish comments you wrote yourself.
Our Author Guidelines ask authors to match the manuscript to CARE and to self-check its length, so you need not count words. We have not published a case-report review form; this guide is advice, not policy. If you find an error in it, please contact us.