Boilerplate Quality Management System for an institution conducting observational studies whose data are intended to be easily integrated into future regulatory submissions by other entities (e.g., as real-world evidence, natural history data, or control-arm comparators).
The QMS consists of a Quality Manual and sixteen SOPs implementing applicable parts of:
Replace the __NewInstitute__ placeholder with your institution's name
and adapt the documents to your structure and risk posture.
- Sized for a new institute running one study. Where a choice existed between "minimal but realistic" and "thorough multi-role separation", the minimal choice was taken. Mature institutes are expected to extend.
- Sponsor-Investigator model. One person holds the combined sponsor and investigator obligations (SOP-001 Section 4.2.2), with a minimal role set and explicitly permitted role combinations and independence constraints (QM-001 Appendix A).
- Inspection posture. The institute is not directly inspected by regulators; the burden moves to data recipients. The QMS therefore emphasizes retrieval, certified copies, and supporting recipient audits rather than hosting regulatory inspections.
- Retention anchored to downstream use, not study end. Canonical retention policy in SOP-002 Section 5.4: default floors of 25 years (data plausibly supporting a marketing application) or 15 years (other regulatory use), never less than 2 years after study discontinuation.
- Single institute-wide logs for nonconformances (SOP-010) and CAPAs (SOP-011), both indexing all intake including items closed without action.
- No per-document review cadence. Document currency is a topic of the annual Management Review (QM-001 Section 5.3.3); no SOP self-reviews.
- Templates summarize; body text governs. Tools and appendices cite the canonical body sections and defer to them on any difference.
- Read-and-attest training. Competence rests on credentials on file, not on testing; training is evidenced by the individual attesting that they read the assigned document, understood it, and had the chance to discuss it with their supervisor (SOP-016). Re-attestation is triggered by substantive revision, not by a calendar.
- Normative statements use bold-caps modals (SHALL, SHOULD, MAY) and nothing else; descriptive prose avoids lowercase modals.
- Cross-references: full markdown link on first introduction, then
"per SOP-00X Section Y.Z" in prose. "Section X.X" in body text; the
section sign (
§) only inside tables. - Each SOP's Section 3.2 lists exactly the documents its body cites. Shared terms have one canonical home and are imported by reference ("as defined in SOP-002 Section 3.3"); a document's body and its own Terms section never reference each other.
- Terms lists are alphabetized (dictionary order: hyphens compare as spaces, spaces significant).
- No em dashes except inside quotations or official titles. Hard wrap at 80 columns (link lines exempt). No single trailing spaces (two or more are a markdown hard break and are allowed).
- Every SOP closes with TRAINING and RECORDS sections; records route to the TMF or quality records per SOP-002.
scripts/qms-check.bash: link, anchor, and style checks encoding the conventions above. Run before committing; exits nonzero on violations.scripts/pandoc-docx-convert.bash: converts all documents to docx indocx/.