Who it’s for

Four people read this report, and they want different things from it.

Including what each one is wary of, which is usually more useful than another list of features.

Safety director

You own the incident program, and you can describe every incident individually while struggling to say anything defensible about the pattern across them.

What you get

  • The conditions that recur across the whole record, each with the population it was counted in.
  • An audit of what your current form can and cannot capture, which is usually the reason the pattern was invisible.
  • Candidate controls written with the observability question attached, so what you would need to check in ninety days is decided now rather than remembered later.

What you’re wary of

That it becomes another dashboard nobody opens. The deliverable is a written analysis and a working session, not a login — and the analysis runs on the record you already have.

Operations

You are asked to change how crews work on the strength of a safety finding, and you carry the cost when the change does not fit the job.

What you get

  • Findings stated at the strength the evidence supports, so you can tell a solid one from a thin one.
  • Explicit evidence limits, including where the record cannot settle the question at all.
  • Controls framed as proposals to test, with what to observe rather than what to assume.

What you’re wary of

That an analysis produces a mandate built on a number nobody can reproduce. Every figure is printed over its own denominator for exactly that reason, and the report says plainly where it stops.

Medical director

You need to know whether clinical and operational conditions are being captured well enough to reason about, and whether anyone is over-reading what the record can support.

What you get

  • A clear statement of which condition domains the reporting instrument can see and which it structurally cannot.
  • Observational language throughout, with recurrence and causation kept apart.
  • Confidence grades that describe evidential support rather than severity.

What you’re wary of

That software will assert causation from co-occurrence. It does not, and it declines to — no finding claims cause from recurrence, and the method page states that boundary in writing.

Fire chief or EMS chief

You have to defend decisions to a board, a union, or a city, and you need evidence that survives being questioned by someone who did not want to hear it.

What you get

  • A traceable document: every claim tied to a population, and the limits stated in the same voice as the findings.
  • A baseline you can measure against later, including when the honest answer is that nothing moved.
  • Nothing that reads as an accusation against a crew member.

What you’re wary of

That it will be used as a disciplinary instrument. Findings are systemic until shown otherwise; the interview is deliberately non-accusatory, and that is a design constraint rather than a policy statement.

Start with the problem, not the product

Five recurring EMS problems, and what a record usually shows about each one — including what it hides.

See the use cases