Stop 2 of 8 · Weekly
Survey notes to findings report
Walkthrough observations become findings with a standard, a risk and a corrective action owner, and nothing is cited until Dr. Salma has the source open in front of her.
The gap between the walk and the report
The walkthrough takes ninety minutes. The report takes most of a day, and almost none of that day is thinking. It is turning eleven scribbled observations into findings, deciding which of them are the same finding, remembering which standard covers reprocessing at this facility type, and building the corrective action table again from scratch.
What changes is that Dr. Salma arrives at the thinking earlier. The grouping, the ranking, the table and the plain English risk lines come back in a few minutes, and the day is spent on the two judgements a consultant is paid for: is this finding real, and is that standard the right one.
What a finding is made of
| Part | What it has to survive |
|---|---|
| Observation | Stated as a count with a denominator, no person in it |
| Standard | Read in the source by you, section and edition confirmed |
| Risk, corrective action, owner, date | A named role at the facility agreeing to a real date |
Severity is the fourth column of your own scale, not Claude's. Dr. Salma uses three levels: immediate patient risk, systemic gap, documentation gap. She tells Claude the definitions and it sorts, but she moves anything that touches sterile processing up a level by reflex.
The verify list
The single habit this stop exists for: Claude never hands back a citation, it hands back a candidate. Every prompt below ends with a section called Verify before citing, and the report does not leave the building with an unticked line on that list. Standards move between editions, appendices get renumbered, and a facility surveyed under one set of conditions is not held to another. Reading the paragraph takes four minutes and is the whole job.
Try this
Below are de-identified observations from a walkthrough at the facility in this Project. Group them into findings, one row per finding, and rank them by risk to patients. Columns: the finding in one sentence, the observations behind it, and the severity from my scale of immediate patient risk, systemic gap, documentation gap. Where two observations are the same underlying failure, say so and merge them. Do not cite any standard yet. [paste the de-identified observations]
For each finding in the table above, name the source and the section you believe applies for this facility type, and say in one line what that section requires. Then repeat every one of them under a heading Verify before citing, written as the exact proposition I am checking, with the document and edition I should open. Mark any where you are unsure whether this facility type is held to that source at all.
Turn the findings into a corrective action table for the facility. Columns: what will change, who owns it by role, and a target that is either within thirty days or within the quarter. Corrective actions must be observable, so write them as something a surveyor could watch happening rather than as a policy being updated. Add one line per finding on how the facility will show sustained compliance at the next visit.
Read the draft findings report as if you were the administrator who wants two of these findings removed. For each finding, give me the strongest objection they will raise and whether it is a fair one. Where the objection is fair, tell me what extra evidence I would need to hold the finding. [paste the draft report]