Could a referral-event QA check distinguish an actual completed handoff from a missing or out-of-order update?
A local synthetic-event validator checks allowed transitions, timestamps and explicit completion evidence while retaining unresolved cases.
What we could measure together
Measure: Referrals with a valid, fully sourced event sequence (percent).
Agree eligible work, observation window, quality threshold and exclusions with the owner before comparing.
Acceptance boundary: No PHI, clinical triage or inferred patient attendance; use fictional IDs.