Skip to content

Living update and release policy

The Wiki is maintained as an updateable research product, but it is not described as a living systematic review. This policy makes currency and revision handling explicit without implying continuous surveillance or exhaustive coverage.

Current release position

Dimension Current position
Evidence search cut-off 22 July 2026
Current retained candidate set 165 candidates: 112 included, 53 excluded
Search status Current pass complete; final pre-submission rerun remains open
Independent review Not complete
Defence validation All candidate requirements unvalidated
Assurance execution All designed tests unexecuted
Response follow-up Attachment provenance is registered; recipient/action/implementation lifecycle remains incomplete

The selection and coverage flow, evidence assurance view and validation queue hold the detailed current state.

Update triggers

An evidence review is triggered when any of the following occurs:

  • a planned submission, presentation or publication is about to be released;
  • a newly published or newly indexed official PFD report may meet the protocol;
  • a new official recipient response affects an included report or a registered counter-position;
  • a primary source is corrected, moved or becomes unavailable;
  • an independent recheck changes inclusion, extraction, coding or causal qualification;
  • Defence workflow evidence, programme specifications or executed assurance results change a REQ-, AST- or gap status; or
  • a material error is reported in the Wiki.

There is no claimed continuous automated surveillance or fixed calendar cadence. At minimum, the targeted search must be rerun immediately before any submission that relies on current corpus counts.

Revision classes

Revision class Examples Required record
Evidence release New inclusion/exclusion, source, response, claim, code, requirement or test status Update canonical records; research log; regenerate; verify; publish with new cut-off
Analytical revision Changed synthesis, controlled mapping or qualification without a new primary source Record rationale and affected IDs in the research log; regenerate and verify
Editorial correction Typography, broken routing or wording that does not alter meaning Technical log or concise correction note; evidence cut-off need not change
Source correction or dispute Official correction, conflicting response, documentary anomaly Preserve prior provenance; add or revise a bounded counter-position; never silently overwrite

Prior states remain recoverable through Git history. The project does not yet claim a DOI-backed archive or formal versioned data repository.

Release gate

Before publishing an evidence-changing release:

  1. update canonical source, screening, included-case, claim and downstream registers as applicable;
  2. record unresolved rechecks in the validation queue and material changes in the research log;
  3. run npm run kb:update, then npm run verify;
  4. inspect the main reader journeys and generated counts;
  5. publish only a clean main commit matching the remote repository; and
  6. state the evidence cut-off and any output-specific older cut-off.

Methodological context: Cochrane living-review guidance, SRC-037 supports explicit surveillance and update-frequency reporting, while PRISMA guidance, SRC-031 supports transparent selection and revision reporting. These sources guide disclosure; they do not confer systematic-review status.