ich-authoring-checklists

Authoring checklists for ICH M4 CTD summaries (M4Q, M4S, M4E), ICH E6(R3) Good Clinical Practice content and ICH E9(R1) estimands and sensitivity analyses. Use when drafting or reviewing Module 2 over

SKILL.md

GitHub

ICH authoring checklists (M4 / E6(R3) / E9(R1))

You draft; the platform verifies. Before citing a guideline revision or asserting a fact is current, ask the connector. Before claiming two documents agree, sweep them.

Start every drafting task the same way

  1. c2c_lookup_ich_guideline for each guideline you will cite (e.g. M4E(R2), E6(R3), E9(R1)) and quote the corpus's title, status and scope.
  2. c2c_check_regulatory_currency with the topic and jurisdiction (and drafted_on when revising an older document) — report facts that are superseded, mandatory-upcoming or stale.
  3. Extract the quantitative claims you will rely on (endpoint results, doses, sample sizes, dates) with their source locations and run c2c_sweep_contradictions before you write. A contradiction blocks the draft until resolved by the author.

M4E — Module 2.5 Clinical Overview checklist

  • 2.5.1 Product development rationale: indication, pharmacological class, regulatory history, scientific advice
  • 2.5.2 Overview of biopharmaceutics: formulation bridging, BA/BE, food effect
  • 2.5.3 Overview of clinical pharmacology: PK, PD, dose–response, special populations, DDI
  • 2.5.4 Overview of efficacy: study design, populations, endpoints, results, subgroups, persistence, limitations
  • 2.5.5 Overview of safety: exposure, AEs by system, SAEs, deaths, labs, special groups, DDIs, withdrawal
  • 2.5.6 Benefits and risks conclusions: therapeutic context, benefits, risks, benefit–risk assessment
  • 2.5.7 Literature references
  • Every number traces to 2.7 or a Module 5 report; run c2c_sweep_contradictions across 2.5/2.7/CSR values

M4Q — Module 2.3 Quality Overall Summary checklist

  • 2.3.S: general information, manufacture, characterisation, control (specification justified against ICH Q6A/Q6B), reference standards, container closure, stability (ICH Q1A(R2))
  • 2.3.P: description/composition, pharmaceutical development (Q8), manufacture, excipients, control, reference standards, container closure, stability
  • Specification acceptance criteria consistent between 2.3 and 3.2; sweep them

E6(R3) — GCP content checklist (protocol / CSR / oversight)

  • Quality by design: critical-to-quality factors identified and risk-proportionate controls described
  • Informed consent process, data protection and participant safety reporting described
  • Investigator, sponsor and IRB/IEC responsibilities and delegation documented
  • Data governance: source, traceability (ALCOA+), computerised systems validation, audit trail
  • Monitoring strategy (risk-based) and its justification
  • Essential records identified and retention stated

E9(R1) — estimand framework checklist (SAP / protocol / CSR)

  • Each primary and key secondary objective has an estimand with all five attributes: treatment, population, variable, intercurrent events with the strategy per event (treatment policy, hypothetical, composite, while-on-treatment, principal stratum), population-level summary
  • Main estimator aligned to the estimand; missing-data assumptions stated
  • Sensitivity analyses target the SAME estimand under different assumptions; supplementary analyses are labelled as such
  • Multiplicity and hierarchy consistent between protocol, SAP and CSR — sweep the alpha allocation and hierarchy statements

Rules

  • Do not state a revision number, effective date or status from memory; the connector's corpus is the citation.
  • Do not compute sample sizes, power or p-values yourself; the platform's deterministic solvers are the source of statistical figures.
  • When the corpus has no entry, say so and cite ich.org for the user to confirm; do not fill the gap.