Medical peer review works best when a reviewer brings clear standards, steady tone, and concrete fixes. This guide lays out a clean way to give a good peer review in medicine that helps editors, lifts authors, and keeps readers safe.
Three anchor resources shape good habits: the COPE ethical guide for reviewers, the ICMJE Recommendations, and the EQUATOR reporting guideline library. Keep them handy while you work.
What A Strong Medical Peer Review Looks Like
A strong report is fair, specific, and ready to action. It starts with a brief, neutral summary, then lists major points, then minor edits, and ends with private notes to the editor. Each point names the exact place in the manuscript, states the issue, and offers a fix the authors can carry out.
| What To Check | What To Look For | What To Write In Comments |
|---|---|---|
| Study question | Clear, clinically relevant question with a defined population and outcome | “State the primary question up front and align it with the stated outcome.” |
| Ethics & approval | Ethics board approval, consent process, trial registration when needed | “Add IRB/REC details and registry ID; clarify consent method.” |
| Study design | Design matches the question (trial, cohort, case-control, cross-sectional, qualitative) | “Explain why this design suits the question; add a diagram of flow.” |
| Participants | Clear eligibility, setting, recruitment, sample size reasoning | “Give inclusion/exclusion lists and how many screened, eligible, and enrolled.” |
| Randomization/blinding | Sequence, concealment, masking of participants, clinicians, assessors | “Describe sequence generation and concealment; state who was masked.” |
| Interventions/exposures | Enough detail to reproduce dose, timing, materials, and adherence | “Provide dose schedules, materials, and adherence measures.” |
| Outcomes | Pre-specified primary and secondary outcomes with valid measures | “Label the primary outcome and timing; justify measurement tools.” |
| Data collection | Standardized procedures, training, and quality checks | “Add data collection SOPs and any calibration or audit steps.” |
| Missing data | Extent, reasons, and handling plan | “Report missingness by group and the handling method (e.g., multiple imputation).” |
| Sample size/power | Plausible assumptions, alpha, and chosen effect size | “Show the calculation and inputs; match them to the primary outcome.” |
| Statistics | Fit of models, checks of assumptions, effect sizes with confidence intervals | “Report effect sizes with CIs; justify model choice and assumption checks.” |
| Results integrity | Totals add up; CONSORT-style flow for trials; no outcome switching | “Add a flow diagram and confirm no changes from the registered plan.” |
| Harms | Adverse events reporting and monitoring | “Include AE definitions, counts by arm, and severity grading.” |
| Tables/figures | Self-contained titles, readable axes, and consistent units | “Retitle tables for clarity and standardize units across figures.” |
| Interpretation | Findings tied to data, care not to overreach | “Trim claims to what the data show; add clinical context and limits.” |
| Generalisability | Who these results likely apply to in real practice | “State which settings and patients this applies to, and where it may not.” |
| Conflicts & funding | Roles of funders and any competing interests | “Clarify funder role and add full conflict disclosures for all authors.” |
| Transparency | Data/code sharing or reasons not to share | “Add a data/code availability statement with links or justified limits.” |
| Reporting checklists | CONSORT, STROBE, PRISMA, SRQR, CARE as suited | “Upload the right checklist and point to where each item sits.” |
| Writing and flow | Plain language, active voice, clean structure | “Tighten long sentences, fix tense drift, and cut duplicate text.” |
Giving A Good Peer Review In Medicine: Step-By-Step
Before You Accept
Scan the title and abstract. Ask: Do I have the skills for the methods, stats, and clinical area? Do I have any ties to the authors or funder? If there is a conflict, decline fast and suggest names. If you accept, note the deadline and set time blocks to work.
First Pass: Big Picture
Read once without marking. Can you restate the main claim and the main result in one or two lines? If not, the paper needs a sharper aim. Start your review with a plain summary in your own words. Editors like to see that you grasp the core of the work.
Check Against Reporting Rules
Pick the right checklist on the EQUATOR site and keep it open. Trials match CONSORT; observational studies match STROBE; reviews match PRISMA; qualitative work uses SRQR; case reports use CARE. Note missing items and ask authors to add them.
Methods And Stats
Look for a clear primary outcome, a prespecified analysis plan, and effect sizes with confidence intervals. Ask for model diagnostics, handling of clustering and repeated measures when present, and a plan for missing data. Request a statistician co-review if the models are outside your lane.
Ethics And Registration
Check for ethics board approval, consent, and trial or systematic review registration where required. Ask for inclusion of registry IDs and any protocol links. If the study poses risk without clear oversight, flag that concern to the editor in the confidential box.
Results And Figures
Do totals match between text, tables, and flow? Are denominators stable across outcomes? Do figures carry units, axes labels, and sample sizes? Ask for raw numbers alongside percentages. Urge authors to lead with absolute risks and patient-centered outcomes.
Interpretation
Ask the authors to state what the data show, what they do not show, and what choices could change the result. Request sensitivity checks and a calm tone on claims. Claims about practice change need a clear link to magnitude, harms, and cost.
Writing Tone
Encourage plain language, short sentences, and direct verbs. Point out jargon and offer a clear rewrite. If the paper is hard to read, list the worst areas and ask for editing by a fluent writer.
Structure Your Report
Use four blocks: a brief summary; major comments; minor comments; confidential notes to the editor. Under each comment, supply the location (page/line/figure), the issue, why it matters for validity or clarity, and a practical fix. Number your points so the authors can reply one by one.
How To Do A High-Quality Medical Peer Review
Tone And Bias Checks
Keep your voice calm and respectful. Critique the work, not the authors. Avoid loaded terms. If you know who wrote the paper, watch for halo effects. Read your own words out loud before sending; trim any snark.
Transparency And Confidentiality
Do not share the manuscript or use its data. State any past ties to the authors, their lab, or the funder. If you talked with a colleague for a second view, disclose that to the editor and add their name if the journal permits.
Template You Can Reuse
Opening summary: one short paragraph in your words.
Major comments: 4–10 items that affect validity or trust.
Minor comments: style, typos, labels, journal style.
Notes to editor: candid view on novelty, risk, and suitability.
What To Say About The Verdict
Give your view on accept, revise, or reject, but let the editor decide. If you ask for a revise, list exactly what changes would move you to a yes. If the flaws are fatal (e.g., wrong design for the question), state that plainly and kindly.
Speed And Timing
Reply to the invite quickly. If life gets busy, tell the editor and propose a new date. Partial reviews hurt more than a fast decline, so bow out early if you cannot deliver.
Reproducibility
Ask for enough detail so another team could repeat the work. Where journal policy allows, ask for data and code sharing or at least a clear path to request them. Encourage authors to move analysis steps to a supplement so readers can follow them.
| Issue | Why It Matters | Sample Comment You Can Reuse |
|---|---|---|
| Unclear primary outcome | Readers cannot tell what was tested | “Please label one primary outcome and move others to secondary.” |
| Outcome switching | Trust erodes when plans change without notice | “Match outcomes to the registered plan or explain any changes.” |
| Under-powered analysis | Risk of false leads rises | “Show the sample size plan and the detectable effect with alpha and beta.” |
| p-values without effect sizes | Readers need size and precision | “Report effect sizes with 95% CIs and de-emphasize p-values.” |
| Confounding not handled | Comparisons may be skewed | “Justify covariates and add a directed acyclic graph in the supplement.” |
| Missing data glossed over | Bias can creep in | “Quantify missingness by group and use a transparent handling plan.” |
| Selective citations | Skips balanced context | “Add recent, balanced citations, including studies with different results.” |
| Claims outrun data | Readers may over-apply the findings | “Dial back claims to match the size, setting, and limits of the study.” |
| Opaque figures | Hard to read means hard to trust | “Use clear axis labels, units, and readable fonts; add data points where fit.” |
| No harms section | Safety picture stays hidden | “Add adverse event methods and counts by group, including severity.” |
Special Study Types
Trials
Look for a prespecified primary outcome, a registry entry, a flow diagram, and masking details. Ask for absolute risk, number needed to treat, and a clear harms table. Point out protocol deviations and how they were handled.
Observational Studies
Ask for a directed acyclic graph or a written causal model, clear handling of time-varying exposures, and checks for collinearity. Urge use of effect sizes that matter to patients, not just models that fit the data.
Systematic Reviews And Meta-Analyses
Ask for a registered protocol, full search strings, risk-of-bias tables, and a plan for small-study effects. Request a study flow diagram and a table that maps inclusion criteria to the final set.
Qualitative Research
Ask for the approach (e.g., grounded theory, phenomenology), sampling, saturation, and coder agreement. Quotes should map to themes, with context on who spoke and where.
Case Reports
Ask for patient consent, de-identification steps, and a clear teaching point. Images need arrows, scale bars, and legends that stand alone.
Ethical Red Flags
Plagiarism, duplicate submission, image manipulation, invented data, or undisclosed conflicts call for a quiet note to the editor, not a public accusation in the author letter. Give specifics and page/figure numbers so the editorial team can act.
When Your Review Triggers A Major Rewrite
When gaps are fixable, point to concrete steps: new analyses aligned to the primary question, tighter outcomes, better figure design, or a clearer write-up. When the design cannot answer the question, say so and advise submission to a venue that fits the work’s scope.
Use The Standards While You Write
When you cite a rule, point to the source. You can reference the ICMJE Recommendations for structure and manuscript basics, the COPE reviewer duties for conduct, and the EQUATOR checklists for study-type specifics. Linking to the exact rule helps authors fix things fast.
Mistakes To Avoid
- Writing a verdict before finishing the methods section.
- Letting tone drift into sarcasm.
- Asking for pet analyses that do not fit the question.
- Copyediting line by line when big design issues remain.
- Ignoring patient-centered outcomes.
Phrases That Keep Authors Moving
- “Please add a one-line aim at the end of the introduction.”
- “State the primary outcome and timing near the start of Methods.”
- “Report effect sizes with 95% CIs; move raw code to a supplement.”
- “Clarify randomization and concealment; name any masking.”
- “Spell out inclusion and exclusion; add a screening flow.”
- “Quantify missing data by group and explain the handling plan.”
- “Shorten the abstract; lead with the main result in absolute terms.”
- “Tone down broad claims and tie them to the data.”
Quick Checks Before You Submit
- Did you add a one-paragraph summary at the top?
- Are major comments grouped by theme and numbered?
- Do minor comments fix clarity, labels, and style only?
- Did you link to COPE, ICMJE, or EQUATOR where you cite a rule?
- Is your tone calm and free of sarcasm or blame?
- Did you declare any ties to authors or funders?
- Is the private note to the editor short and candid?
How To Close Your Report
End with a short list of the most useful fixes from your review. Thank the authors for the chance to read their work. State your availability to re-review after a careful revision. Save a copy for records.