| name | clinical-submission |
| description | Write a clinical study report on the ICH E3 skeleton — Synopsis, Ethics, Investigators/Structure, Objectives, Investigational Plan, Methods (efficacy & safety), Results, Discussion & Conclusions, Tables/Figures/Appendices — situated in the CTD five-module frame (M1-M5). Use when the deliverable must reproduce the clinical-study-report structure of a regulatory submission. Anti-trigger; for a scientific journal article write academic, for a financial/legal disclosure filing use regulatory-disclosure, not this genre — and this genre never asserts clinical validity or regulatory acceptance, only structure. |
| argument-hint | <the clinical study to report on> |
clinical-submission
Produces a clinical study report (CSR) on the ICH E3 skeleton, situated
in the CTD (Common Technical Document) five-module frame. Its center of
gravity is the fixed section order — Synopsis through Tables/Figures/
Appendices — with efficacy and safety kept distinct throughout, and every claim
resolving to a cited finding. This genre follows the ICH E3 clinical-study-report
convention used across regulated submissions.
Scope caveat (carry, do not over-sell): this genre reproduces the ICH E3
clinical study report structure within the CTD module frame. It does not
assert clinical validity, statistical adequacy, or regulatory acceptance — the
analysis is only as sound as the cited findings. Do not market output as a
"submittable" CSR.
Pattern (industry: ICH E3 CSR, CTD module frame)
- Synopsis — a structured summary of the study and its principal results.
- Ethics — ethics-committee review, informed consent, and conduct standards.
- Investigators & Study Structure — administrative structure and
investigators.
- Objectives — the primary and secondary study objectives.
- Investigational Plan — study design, randomization/blinding, and
rationale.
- Methods (Efficacy & Safety) — endpoints, analysis populations, and
statistical methods, kept distinct for efficacy and safety.
- Results — efficacy results then safety results, with the analysis
populations.
- Discussion & Conclusions — interpretation, limitations, and
benefit-risk.
- Tables, Figures & Appendices — the supporting data displays and
appendices.
CTD framing: situate the report in the five-module Common Technical
Document frame — M1 regional administrative, M2 summaries, M3
quality, M4 nonclinical study reports, M5 clinical study reports — and
state that an E3 CSR lives in Module 5.
Rules that keep it a clinical study report
- Every claim is traceable to a cited source; no orphan facts. Verification
verdicts are reported, not silently dropped — annotate
weakened or
inconclusive findings with their uncertainty qualifier, and exclude only
falsified findings.
- State limitations and benefit-risk honestly; the genre reproduces structure,
not clinical validity or regulatory acceptance — say so. An undiscussed
safety signal is a defect.
- Efficacy and safety stay distinct — in Methods, in Results, and in the
Discussion's benefit-risk weighing. Never merge them into one undifferentiated
"results" narrative.
- Hedge uncertain claims; present confidence intervals or ranges when sources
disagree.
- Exhaustive coverage: build the report from the full surviving evidence —
every surviving finding is treated with its own evidence (claim, citation,
population), never condensed to a cherry-picked subset. A silently dropped
finding is a defect.
- Section order is fixed — Objectives precedes Investigational Plan precedes
Methods precedes Results. A CSR that reorders or omits a required section is
not conformant.
- Verify live: check the current ICH E3 guidance and CTD module numbering at
authoring time rather than baking a guidance revision into output as settled
fact — regulatory guidance revises.
Required Figures & Matter
- Front matter: title page identifying the study, and the Synopsis.
- Figures: efficacy and safety tables (analysis populations, endpoint
results, adverse-event summaries); number and caption each and reference it
in the text. Any efficacy or survival curve is rendered as a Mermaid
xychart-beta. Any figure, chart, or diagram is a fenced mermaid code
block — never ASCII art, an image link, or Graphviz/DOT — and a required
figure is never silently omitted; if the data cannot support it, say so in
prose. Plain tabular matter stays a Markdown table.
- Back matter: appendices and the full reference list.
Citation Style
Scientific / regulatory referencing to source studies and guidance. Every claim
resolves to a source with a URL (MIF Level 3 floor); no uncited claims.
FDA eCTD v4.0 electronic packaging is an orthogonal serialization — it is out
of scope for this genre.
Anti-triggers — do not use this genre for
- A scientific journal article reporting the same trial for peer review —
that is
academic: discipline-citation-style narrative, not the fixed ICH E3
CSR skeleton or CTD module frame.
- A financial or legal disclosure filing — that is
regulatory-disclosure
or a similar filings genre: disclosure obligations under securities or
contract law, not a clinical-trial write-up.
- The electronic packaging of a submission (eCTD v4.0) — that is a
serialization/transport concern, a separate
ectd-style channel, not this
document genre.
MIF frontmatter
type: semantic — a clinical study report is declarative study-and-results
knowledge, not a time-bound event or a step sequence. Climb to L2 with
namespace (clinical-submission/<area>), modified, title, and tags when
the review context supplies them. At L3, entity.entity_type is
clinical-study-report. Gate every output with mif-validate at its target
level; the floor is --level 1.
Why machine-readable — the point of MIF here
| Question an agent asks | Answered by (frontmatter) |
|---|
| Is this report still the current version of the study record? | temporal.validFrom / ttl |
| Where did each result come from; can I trust it? | provenance (W3C-PROV) + citations[] |
| What does this CSR relate to (protocol, statistical analysis plan)? | typed relationships[] (relates-to) |
| Which results are confirmed vs. exploratory/underpowered? | citations[] tied to each Results claim, with verdict annotated in prose |
The same document still reads as a human clinical study report and projects
losslessly to JSON-LD and back — one artifact, two readers.
The L1 -> L3 climb (two exemplars)
templates/good-l1.md — L1 floor: id, type, created + body. A
complete, valid CSR, but opaque to a machine consumer.templates/good.md — L3 (highest this genre supports): adds
namespace, modified, temporal validity, W3C-PROV provenance,
citations[] tied to the efficacy/safety claims, and a typed
relationships[] graph (e.g. relates-to the statistical analysis plan).
Validate with mif-validate --level 3.
Author at the highest level the drafting context supports (grade down
rather than fabricate). templates/bad.md shows the antipattern: a report
that collapses efficacy and safety into one undifferentiated narrative in
Methods and Results instead of keeping them distinct throughout.
