From lancet-skills
Ensures clinical trial design meets Lancet standards: prospective registration, protocol/SAP, and design rigor for RCTs and observational studies.
How this skill is triggered — by the user, by Claude, or both
Slash command
/lancet-skills:lancet-study-designThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
- Planning or writing up a clinical trial or a major observational study.
The Lancet follows ICMJE: clinical trials must be registered prospectively — before the first participant is enrolled — in a WHO-ICTRP primary registry or ClinicalTrials.gov.
| Registry | Use / region |
|---|---|
| ClinicalTrials.gov | US and international |
| ISRCTN | UK / international |
| WHO ICTRP | global portal aggregating primary registries (e.g., EU-CTR, ANZCTR, CTRI, ChiCTR, PACTR) |
The Lancet often handles large international / multi-country trials — confirm registration in each relevant jurisdiction and report the lead registry.
Registration timing is the most consequential design fact The Lancet checks: following ICMJE, it will decline to consider a trial not registered prospectively, and an outcome switched from the registered one is an integrity concern. Desk-reject or major-revision triggers: retrospective registration left undisclosed; a switched primary outcome; no protocol or pre-specified SAP; concealment conflated with blinding; per-protocol presented as primary for a superiority trial. Confirm any current protocol-submission requirement against the journal's author guidelines.
A hypothetical double-blind, multi-country superiority RCT.
Design lock (illustrative):
Registry ISRCTN registered 2021-03 BEFORE first enrolment 2021-06 -> prospective, PASS
Registered primary outcome = reported primary outcome -> no switch, PASS
Randomisation 1:1 stratified by country; central web concealment; double-blind
Power: ARR 6 pp, alpha 0.05, power 90% -> n=2 040 (1 020/arm), +8% attrition
Primary population ITT; per-protocol pre-specified sensitivity; independent DSMC
The design passes: registration is prospective and dated, outcomes agree, concealment and blinding are separate, and the sample size is justified.
【Study type】 RCT (parallel/cluster/crossover/factorial/adaptive/NI) / observational (cohort/case-control/cross-sectional)
【Registration】 registry + number + prospective? (before first enrollment: yes/no) + in abstract? yes/no
【Protocol + SAP】 present? primary analysis & populations pre-specified? yes/no
【RCT rigor】 randomization / allocation concealment / blinding / primary endpoint / power / ITT → gaps
【Observational rigor】 design / confounding strategy / causal language calibrated → gaps
【Deviations from protocol/SAP】 [...] reported transparently? yes/no
【Next】 lancet-reporting
npx claudepluginhub brycewang-stanford/awesome-journal-skills --plugin lancet-skillsConfirms study-design rigor and mandatory prospective trial registration, protocol, and statistical analysis plan before writing up a clinical study for NEJM.
Guides study design and internal-validity safeguards for JAMA clinical manuscripts (RCT, cohort, diagnostic, systematic review). Covers intention-to-treat, blinding, confounding control, and design-checklist alignment.
Final preflight checklist for Lancet submissions covering importance, registration, reporting guidelines, statistics, abstract, ethical declarations, and required files. Bundles a cover-letter template.