Durable Institutional Memory

The θ That Survives Staff Rotation

The prior update from Session 1 is only useful if it survives. Staff rotate. Institutional memory evaporates. This session is about the infrastructure that keeps θ′ alive and accumulating across personnel change.

The rotation problem

A WHO country office team that produces excellent evidence briefs but does not document its priors is running a Markov process with no memory. Each new staff member starts from θ₀ — the same uninformed prior that the team started from three years ago. Every hard-won insight about which PECO-F framings work for India health financing, which sources are reliable, which anchor populations are systematically missing from the literature — all of it exits with the departing staff member.

Durable institutional memory requires four concrete artefacts, each of which externalises a component of the team's accumulated θ into a format that survives personnel change:

The prior library

A shared, version-controlled document containing: the team's accumulated PECO-F templates by topic area, the most productive Boolean strings, the known ε gaps by population, and the anchor constraint definitions. Updated after every brief. Stored in a shared drive, not a personal folder.

The brief registry

A searchable log of every brief produced: topic, date, key findings, evidence gaps named, recommendation, Ministry response, and follow-up status. Two pages maximum per entry. New staff can read backward through the registry to reconstruct the team's evidence trajectory without re-reading full briefs.

The SOP stack

The handover-ready SOPs built in Level III Session 2, maintained and updated after each major brief cycle. One SOP per workflow type. A new team member can run the workflow from the SOP alone — this is the test of whether the SOP is sufficient.

The onboarding brief

A single two-page document written for each new team member: where the evidence base currently stands, what the active surveillance topics are, what the anchor populations are for current briefs, and where the known ε gaps are. This is the compressed θ — the current prior handed to a new arrival on day one.

The 20-minute maintenance habit

None of these four artefacts requires significant time to maintain if updated consistently. The failure mode is letting updates accumulate until the task feels overwhelming — at which point it is never done. The discipline is: 20 minutes after every brief is circulated, update the prior library (one new PECO-F template or ε gap), add a brief registry entry (two paragraphs), and flag any SOP updates triggered by the brief's production. That is the entire maintenance cost of durable institutional memory.

🎯 Key takeaway

Institutional memory is not what people remember — it is what has been written down and kept somewhere the next person can find it. Four artefacts make θ′ durable: the prior library, the brief registry, the SOP stack, and the onboarding brief. Twenty minutes after each brief. That is the cost of a learning institution. Session 3 addresses the other side of durability: keeping the prior current through continuous surveillance rather than letting it age between reviews.