Peer review explained: from editorial screening to a useful revision

A reviewer report is useful when it leads to a clear scientific question, a documented check and a response the editor can evaluate. Authors can make that process easier by separating the manuscript’s current state from the changes they propose. This guide follows that distinction through screening, review and revision, with a worked teaching example.

1. Identify the decision actually in front of you

An editorial office may first ask for missing files or a formatting correction. That request does not establish that external review has begun. A journal can also decline a paper without sending it to reviewers. ICMJE describes peer review as expert advice that supports editorial judgment; the editor remains responsible for the selection of content and need not follow a reviewer’s recommendation.

Keep a simple correspondence record: manuscript identifier, date, exact status or request, responsible author and next action. If a message only confirms receipt, record receipt. If the editor requests a revision, preserve the decision letter with the version it concerns. This prevents a team member from treating a portal label or a favourable comment as acceptance.

Reference: ICMJE: submission and peer-review responsibilities

2. Read the journal’s review policy before preparing files

Review arrangements differ across journals, including masking of identities and whether reports are published. Read the chosen journal’s policy for the article type you are submitting. Check the main file, acknowledgments, file properties and supplements against its instructions; removing names from the title page alone may not prepare a file for anonymous review.

A published schedule can help you plan a follow-up, but your manuscript’s actual progress depends on the editorial process. The transparency principles reject guaranteed acceptance and require review-time claims to have supporting published evidence. A payment, language edit or automated preparation report should not be treated as a decision on scientific merit.

Reference: COPE / DOAJ / OASPA / WAME: transparency principles

3. What makes a reviewer’s comment actionable?

A useful comment identifies the issue, points to the relevant passage or result and explains why it affects interpretation. “The comparison needs a clearer denominator” gives an author a task. A personal judgment about the authors does not. Separate major issues affecting the study’s interpretation from minor presentation corrections, and flag when a concern lies outside your expertise.

For invited reviewers, follow the journal’s confidentiality and conflict policies. ICMJE’s AI guidance requires adherence to the journal’s AI policy or prior permission, protects manuscript confidentiality and calls for disclosure and checking of AI-assisted review. Possessing a manuscript file does not by itself permit uploading it to an external tool.

Reference: ICMJE: AI use by reviewers

4. Build a response matrix before rewriting the letter

Read the editor’s instructions first, then give each reviewer point its own row. Assign the scientific check to the person who can verify it. Use separate states for proposed work and completed work. Add the location only after the revised file exists, and check that location again after final pagination.

The example below is entirely fictional. Assume that the example’s author team has checked the revised methods, table, abstract and limitations wording. The third request remains unsupported by its dataset; the response makes that limitation visible.

On small screens, scroll the table horizontally, or focus it and use the arrow keys.

Fictional revision matrix: verified work and honest limits
Reviewer point	Verified state in this example	Suggested response
Explain which records were eligible.	Eligibility wording added in Methods, paragraph 2.	We clarified the eligibility rules in Methods, paragraph 2. The revised text identifies the records included in the review.
State the denominator for completeness.	Table 1 now reports 90 complete records out of 120 assessed; the same figures were checked in the abstract.	We now report the count and denominator as 90/120 (75%) in Table 1. We checked the same figures in the abstract.
Show whether documentation improved patient outcomes.	Patient outcomes were not collected; no new analysis performed. The limitations wording was checked and contains no patient-benefit claim.	The dataset does not contain patient outcomes, so we cannot test this question. We identify this limitation and avoid a claim of patient benefit.

5. Disagree precisely, and ask for clarification when needed

An author can explain why a requested change would answer a different question, require unavailable data or introduce an inappropriate analysis. Describe the limitation, the evidence and the wording you propose. Avoid blanket statements that a reviewer is wrong. If two requests conflict, show the editor the exact conflict and ask which approach to prioritise.

When a deadline cannot be met, request an extension through the journal’s stated channel with a concrete account of the outstanding work. For an appeal, read the journal’s policy and distinguish a demonstrable factual or procedural issue from disappointment with the outcome. Preserve relevant correspondence privately and do not disclose confidential review material in a public discussion.

6. Verify the revision package and the final handoff

Open the files you intend to submit. Check the clean manuscript, marked version if requested, response letter, figures and supplements against one another. Search for superseded numbers, unresolved comments and placeholder citations. Ask each responsible co-author to approve the final package. Under ICMJE authorship criteria, final approval and accountability accompany substantive contribution and drafting or critical review.

After submission, save the actual receipt and package version. A further review round may follow. If the editor later accepts the paper, examine the proof for changes introduced during production, including author details, symbols and reference links. The scholarly record still needs attention after publication: a reader’s report of an error deserves a documented assessment through the journal’s correction process.

Reference: ICMJE: authors and contributors

Questions for a practical final check

Does each reviewer point have a response and a verifiable manuscript location where relevant?
Are all claims of completed analysis supported by actual outputs?
Do the abstract, tables and response use the same denominators?
Have confidential materials remained within the journal’s permitted workflow?
Have co-authors approved the exact files being resubmitted?
Does the saved receipt identify the submitted version?

Examples illustrate a working method; they do not establish the readiness of a particular journal or manuscript. This is general guidance: check the current policies of the journal concerned. Source links reviewed 14 September 2026.