Journal of Ophthalmic Science · Peer-review practice
Guidelines for Reviewers
A JOS review should help the editor judge whether an ophthalmic manuscript is valid, ethically reported, clinically interpretable and supported by its evidence. Reviewers advise; editors remain responsible for the decision.
Before accepting
- The clinical subject and methods fall within your competence.
- No relationship or interest could reasonably compromise objectivity.
- You can complete a confidential review by the date in the invitation.
I
Accepting or declining an invitation
Expertise
Accept only when you can assess the manuscript’s principal ophthalmic question and methods. Tell the editor which aspect falls outside your competence so complementary expertise can be added.
Conflicts
Disclose personal, financial, professional, institutional, supervisory, competitive and recent collaborative relationships. When uncertain, disclose and allow the editor to decide.
Availability
Use the deadline stated in the invitation. Decline promptly if it cannot be met, or contact the editor before accepting if clarification is required.
Confidentiality begins with the invitation
Do not share the manuscript, figures, data, author information or invitation with colleagues or trainees without prior editorial permission. Do not contact the authors directly or use unpublished content for personal, clinical, commercial or research advantage.
II
A structured assessment
The review should test the study against its own question and design. Novelty, statistical significance or a striking image cannot compensate for invalid methods, missing ethics or conclusions unsupported by the evidence.
| Criterion | Questions for the reviewer | Useful report content |
|---|---|---|
| Scope and question | Is the primary contribution ophthalmic? Is the question specific, answerable and important in its clinical or translational context? | State the manuscript’s claimed contribution and whether the study actually addresses it. |
| Design and population | Does the design fit the question? Are eligibility, setting, sample, comparators, follow-up and sources of bias described? | Identify design limitations that alter interpretation and distinguish remediable reporting gaps from unresolvable flaws. |
| Analysis | Are outcomes prespecified and clinically meaningful? Are effect estimates, uncertainty, missing data and multiplicity handled appropriately? | Explain how any analytical issue affects a result or conclusion; do not request significance alone. |
| Results and claims | Do tables, figures and text agree? Are benefits, harms, diagnostic performance and limitations reported without selective emphasis? | Point to the exact result that supports or contradicts each central claim. |
| Reporting | Is the relevant reporting guideline followed and is the work reproducible or clinically interpretable from the information supplied? | Request only information required to assess or reproduce the work, and name the applicable standard where useful. |
| References | Is the literature current, balanced and directly relevant? Are important contrary findings represented? | Suggest citations only when they materially correct context; never require citation of the reviewer’s work for personal benefit. |
III
Ophthalmic and clinical checks
Human participants and trials
Check ethics approval, informed consent, Declaration of Helsinki reporting and prospective public registration for applicable interventional studies. Compare registered and reported outcomes where possible.
Patient privacy
Identify recognizable periocular or facial images, rare-case combinations or clinical detail that may identify a patient. A bar over the eyes is not a substitute for written publication consent.
Imaging and image integrity
Assess acquisition settings, scale, orientation, grading, adjudication, quality control and whether adjustments were applied uniformly. AI-assisted enhancement, segmentation or analysis should be disclosed.
Diagnostic accuracy
Evaluate participant selection, index test, reference standard, blinding, indeterminate results, threshold selection, sensitivity, specificity and uncertainty using the STARD framework where applicable.
Surgery and treatment
Check intervention detail, operator context, comparator, learning curve, follow-up, protocol deviations, complications and adverse events. Claims of benefit should be proportionate to the design.
Devices and pharmaceuticals
Assess sponsor involvement, author relationships, comparator choice and analytical independence. Undisclosed commercial influence or missing safety information should be raised confidentially with the editor.
Case reports and image reports
Look for a clear clinical learning point, adequate diagnostic reasoning, relevant differential diagnosis, outcome, appropriate consent and adherence to CARE where applicable.
Data transparency
Check that the data-availability statement is informative and consistent with the study. Flag implausible values, unexplained exclusions or evidence that cannot support the reported analyses.
IV
Writing the review report
- Summarize the manuscript fairly. State the question, design and main contribution in a few sentences so the editor can see how you interpreted it.
- Identify the decisive issues. Put validity, ethics, interpretation and missing evidence before stylistic or presentational points.
- Separate major and minor comments. Explain why each major issue matters and what would be required to address it.
- Be specific and proportionate. Refer to the relevant section, table or figure and avoid requesting work unrelated to the research question.
- Review the work, not the authors. Use professional, objective language without sarcasm, speculation about motives or discriminatory assumptions.
Comments to authors and editor
Place scientific comments that authors can address in the author-facing report. Do not conceal a decisive methodological objection only in confidential comments.
Use confidential comments for the recommendation, suspected misconduct, identity-sensitive conflict information or concerns that should first be assessed by the editor.
V
Recommendation categories
Choose the category that matches the scientific issues described in the report. The recommendation does not determine the editorial decision.
- Accept
- The evidence and reporting are complete; any remaining changes are production-related.
- Minor revision
- Bounded clarifications or corrections are required, but the central design, analysis and conclusions remain sound.
- Major revision
- Substantial but feasible analysis, reporting or interpretive changes are required before the manuscript can be reassessed.
- Reject
- The manuscript is outside scope or has design, ethics, integrity or evidentiary problems that cannot be resolved within the present study.
VI
Revised manuscripts, suspected misconduct and AI tools
- For a revised manuscript, compare the response letter with the revised text and confirm whether each substantive concern was addressed.
- Raise a new issue only when it is material and could not reasonably have been identified in the previous round.
- Report suspected plagiarism, duplicate publication, fabrication, image manipulation, undisclosed conflicts or authorship problems confidentially to the editor. Do not investigate or contact authors independently.
- If a conflict becomes apparent after accepting, stop the review and inform the editor before proceeding.
- Submit the report by the invitation deadline. If delay becomes unavoidable, contact the editor so the manuscript can be managed without misleading the authors.
AI and manuscript confidentiality
Do not upload manuscript text, images, data or review material to a public or unauthorized AI service. Such disclosure can breach confidentiality, privacy and intellectual-property obligations.
An automated system must not replace the reviewer’s expert judgment. Any authorized tool use must comply with JOS policy and be disclosed to the editor.
After completing the review, securely delete any downloaded manuscript files, images and local notes unless the journal has instructed you to retain them within an authorized system.
Questions about an invitation or a confidential concern may be sent to [email protected] with the manuscript identifier and no unnecessary patient information.
Complete an invited review
Use the secure manuscript system to accept or decline the invitation, read the files and submit confidential and author-facing comments.
Open ManuscriptZone