lancet-rebuttal
GitHub用于处理柳叶刀审稿意见，包括编辑决定、统计及临床评论。通过四象限分类策略，指导用户谨慎修改因果表述，提供包含引用、证据和修订文本的点对点回复信草稿，确保符合期刊规范。
Trigger Scenarios
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill lancet-rebuttal -g -y
SKILL.md
Frontmatter
{
"name": "lancet-rebuttal",
"description": "Use after Lancet reviews arrive (including the statistical reviewer and the editor's letter) to triage the decision, answer statistical and clinical comments, and draft a point-by-point response that quotes the reviewer, gives evidence, quotes the revised manuscript text, and keeps causal language cautious. Do not run before the main text is actually revised."
}
Reviewer Rebuttal (lancet-rebuttal)
When to trigger
- A decision letter arrives (reject / major or minor revision).
- You have clinical and statistical reviewer comments and need a strategy before writing.
- A revision is drafted and you need the point-by-point response letter.
Step 0: read the editor's letter first
The editor's framing outranks individual reviewers. Identify:
- The decision type and whether a new review round is implied.
- Which concerns the editor emphasises (load-bearing — address fully).
- Any deal-breaker the editor names (e.g., "the primary analysis must follow the pre-specified SAP," "registration timing must be clarified," "the causal claim must be softened").
A "we would consider a revised manuscript" is an invitation gated on the deal-breaker. Don't treat it as a flat reject, but don't under-deliver.
The statistical reviewer is its own track
The Lancet routinely uses a statistical reviewer. Treat their comments as a distinct, high-priority track:
- Answer with the pre-specified analysis and SAP references; show CIs and absolute effects.
- If asked to add an analysis, add it and label pre-specified vs post-hoc honestly.
- Address ITT/per-protocol, missing-data handling, multiplicity, and subgroup interaction concerns with actual re-analysis, not prose.
Triage every comment into 4 buckets
| Bucket | Action |
|---|---|
| Do (fair, feasible) | Make the change; show it; quote the new text/table/figure. |
| Do-partial | Do what is feasible; explain the boundary with evidence. |
| Defend (incorrect/out of scope) | Push back respectfully, with data/citations — not assertion. |
| Defer (future work) | Acknowledge; add a sentence to the text; don't over-promise. |
Most rejections-on-revision come from silently skipping a load-bearing comment or defending where a re-analysis was actually needed.
Response-letter format
For each comment:
Reviewer N (or Statistical Reviewer), Comment k: <verbatim quote of the reviewer>
Response: <what we did / our reasoning>. <Evidence: new table/figure, re-analysis,
effect with 95% CI.>
Changes: "<quoted new manuscript text>" (p. X, lines Y–Z; new table/figure/appendix).
- Open with a short thank-you and a 3–4 line summary of the major changes across the revision.
- Quote each comment verbatim; never paraphrase in a way that softens the reviewer's point.
- Quote the new manuscript text so the editor can verify without hunting.
- Use a consistent visual convention (reviewer text plain; response indented; manuscript quotes in italics/colour).
Tone and causal-language rules (Lancet-specific)
- Respectful and non-defensive; if a reviewer misread the paper, the writing was unclear — fix the writing and explain.
- Keep causal language cautious: if a reviewer says the conclusion outruns the design (e.g., causal claim from observational data, or a superiority claim from a non-inferiority trial), this is the most dangerous comment — either add evidence or narrow the claim in the title, abstract Interpretation, and Discussion, and say you did.
- Concede real limitations explicitly; an honest limitations paragraph builds credibility.
Update the Research in context panel if the evidence base changed
If new studies appeared during review, re-run or update the systematic search and revise "Evidence before this study" and "Implications of all the available evidence" accordingly — and note this in the response (see lancet-research-in-context).
The Lancet-specific reviewer comments and how to land them
The comments that decide a Lancet revision cluster around the venue's identity: registration integrity, the statistical reviewer's re-analysis requests, claim calibration, and the global-health/equity framing. The fix for each is concrete, never rhetorical — the editor verifies it against quoted new text.
| Reviewer comment | The venue-specific fix |
|---|---|
| "Trial not prospectively registered." | Disclose exact registration/enrolment dates; if retrospective, flag it and explain — do not bury it |
| "Primary outcome switched." | Reconcile to the registered outcome, or document the dated change; label post-hoc analyses |
| "Causal claim outruns the design." | Narrow the title, abstract Interpretation, and Discussion; say you narrowed it |
| "Global-health relevance / equity not addressed." | Add the LMIC/PROGRESS-Plus framing in Interpretation; coordinate with lancet-fit |
| "Statistical reviewer: add ITT/sensitivity analysis." | Run the re-analysis; report effect + 95% CI; show it as a new table, not prose |
| "Reporting checklist incomplete." | Complete and page/line-map the EQUATOR checklist; resubmit as a supplement |
Worked micro-example (illustrative — not a real exchange)
A statistical-reviewer comment on a hypothetical superiority RCT, answered the Lancet way.
Statistical Reviewer, Comment 3: "The primary analysis is reported as per-protocol.
For a superiority trial, the intention-to-treat population should lead."
Response: We agree. We have made the pre-specified ITT analysis primary, with
per-protocol as a labelled sensitivity analysis. The conclusion is unchanged.
Evidence (illustrative): ITT response 60.0% vs 50.2%, absolute risk difference
9.8 pp (95% CI 5.4-14.2; p=0.0003); per-protocol 62.1% vs 50.0% (RD 12.1 pp,
95% CI 7.3-16.9) -> consistent.
Changes: "The primary analysis was by intention to treat..." (p. 8, lines 12-19;
new Table 2 row; appendix table S4 for per-protocol).
The request is answered with an actual re-analysis carrying effects and 95% CIs, the ITT/per-protocol order is corrected, and the new manuscript text is quoted so the editor can verify without hunting — the pattern that converts a statistical-reviewer comment into an accept.
Output format
【Decision type】 reject / major / minor
【Editor's load-bearing concerns】 [...]
【Deal-breaker】 ... → plan to resolve
【Statistical-reviewer track】 [...] → re-analyses planned
【Comment triage】 Do [...] / Do-partial [...] / Defend [...] / Defer [...]
【Claim integrity】 any narrowing of causal/practice claims needed? (re-check lancet-fit)
【Panel update】 systematic search re-run? Evidence-before / Implications revised? yes/no
【Response letter】 drafted point-by-point with quoted changes
Anti-patterns
- Do not silently skip a comment the editor emphasised.
- Do not answer a statistical-reviewer request for re-analysis with prose alone.
- Do not defend an over-reaching causal claim — narrow it.
- Do not draft the response before the manuscript is actually revised.
Version History
- 1839142 Current 2026-07-05 14:01
