The problem is not always the record
Care teams already document a lot. Incidents get recorded. Visits get filed. Discharge notes land somewhere. Supervisors review what they can. The failure is often not that nobody wrote anything down.
The harder problem is that the useful context gets spread across systems, shifts, inboxes, handoffs, and memory. Each individual record may be accurate, while the pattern connecting those records remains hard to see.
A familiar kind of gap
A home health agency operating 14 active patient routes receives two insurance authorization denials in the same month. Both involve patients discharged with port-a-cath orders requiring home infusion. In each case, the order was missing the co-signature required before the first scheduled visit.
No system flags the pattern. One person notices it because she remembers the same kind of problem from last month.
“Mr. Alvarez was re-admitted Monday. His port-a-cath wasn't co-signed before the home infusion visit and insurance denied it. This happened last month too, with a different patient from the same unit. Someone needs to own a protocol for BIDMC discharges with port-a-cath. It's a gap and it's going to keep biting us.”
What gets lost between the records
A system of record can store the denial, referral, visit, and discharge note. But the operating truth lives between those records: the recurring source, the missing co-signature pattern, the staff member who noticed it, and the work still required to prevent it from happening again.
The real question is not only, “Was this documented?” It is, “Will the team remember what this taught us when the next case appears?”
This is what we mean by operating memory
Operating memory is the layer of working knowledge that helps a care organization carry context forward: what changed, what keeps recurring, what remains unresolved, who owns the next step, and what evidence supports the decision.
Verity starts from the way that knowledge naturally appears: a voice note, handoff, family update, exception, incident, visit note, or supervisor correction.
Protocol update
BIDMC discharge, port-a-cath, home infusion authorization
NP must co-sign home infusion order before first patient visit
Two insurance authorization denials in 30 days; operational risk
Boston South Care Team
Before activation, the draft goes to a designated reviewer. The reviewer confirms scope, corrects language, and decides what becomes official. AI drafts, humans approve, and operators retain accountability.
The memory has to become work
The approved record does not just sit in a file. It creates open loops with owners, due dates, and source evidence, so the organization can see what remains unresolved.
Previous denial — confirm insurance resolved
Due May 28Notify BIDMC discharge team of new protocol
Due Jun 1What the team gets back
Once approved, the pattern becomes accessible to the people who are allowed to use it: the next shift, the supervisor, the route coordinator, and later the manager asking why denial burden increased.
Include co-signature check for relevant discharges.
Track denial resolution and discharge-team notification.
Answer only from policy-eligible approved memory with source links.
Surface recurring authorization issues as continuity signals for human review.
The takeaway
Frontline care teams do not only need another place to type notes. They need a practical way to carry forward the context that records alone do not keep alive.
That is where operating memory becomes useful: capture what happened, review what becomes record, route what needs doing, and preserve approved context for the next shift, the next supervisor, and the next decision.