SurvAIHealth
While the details are still fresh

Find the gap while the people who were there still remember it.

Every incident is read against the F-tags a surveyor would use, with your own chart quoted back. Anything thin surfaces the same day, so the staff who were on the floor can complete the record properly — instead of somebody reconstructing it from memory a year later, under a surveyor's question.

Seven aspects, each a real question

An incident arrives from PointClickCare and the relevant worksheets are worked: the assessment that should have followed, the care plan revision, the notification, the intervention and its review, the supervision in place beforehand, and the environment.

Each row is answered from the record — the note, the order, the assessment index, the incident itself — and the answer carries the evidence it rests on, with dates.

Four states, because a clean row and an unread row are not the same

StateMeans
MetWe read the chart and the documentation is there. The quote is shown.
Not metWe read the chart and it is not there. This is a finding.
UnknownWe read the chart and cannot tell. Not a finding, and never counted as one.
Could not lookA record we do not have access to. Never shown as a clean result.

Most audit tools have two of these. Collapsing the middle two into either of the outer ones is how a compliance tally quietly stops meaning anything — and it is the single defect this product has spent the most effort removing.

A name we do not recognise is not an absence

Every facility names its own forms and every pharmacy its own products. So when we look for a post-fall assessment and find nothing under that name, there are two possible explanations: it does not exist, or it exists under a name we do not know.

Only the first is a finding. So that row comes back as unknown, and it lists the assessments that were on file around that date — which is the half a reviewer can actually act on. Writing it as "not met" would have been easy, would have looked more decisive, and would have been wrong often enough to matter.

Things we cannot see, named

These are not coming in a later release, and we would rather you knew now than found out during a survey:

That last one is why there is no column for it. A blank field for a fact we cannot read is how "nothing recorded" starts being read as "no dose given".

Restraints, and an open question we have not closed

The restraint review checks the three things F604 turns on: the device, the medical symptom warranting it, and the instruction to release and reposition. Where an order matches, it reports against all three.

Where no order matches, it returns unknown and lists the orders that are on file — because whether a siderail or a lap belt surfaces on the medication and order feed at all, in a live tenant, is a question we have asked PointClickCare and have not yet had answered. Until we have, a no-match says nothing about your building, and the row says exactly that.

Quoted from the chart Unknown is not a finding Blind spots named Runs on arrival

Related: The 24-hour chart check · Drafting the plan of correction · What it reads from PointClickCare

Run it on an incident you have already closed.

You know what that chart holds. It is the fastest way to find out whether we do.

Start the 60-day pilot