- A single six-month records request typically consumes two to four staff days at a mid-size program.
- Almost none of that time is clinical judgement. It is exporting, collating, chasing signatures, and numbering pages.
- The hours are invisible because they are absorbed by salaried people who were going to be paid anyway.
- The cost that matters is not the hours. It is what those people stopped doing, and what gets missed under deadline.
Ask a program what their last records request cost and you will get a shrug, because nobody bills for it and nobody tracks it. So let us count it.
The work, step by step
A payer requests six months of records for one patient in a residential program. That patient has an assessment, a treatment plan with three reviews, forty-eight individual notes, a hundred and twenty-six group notes, eight level-of-care reviews, twenty-two assessment results, and a folder of authorization correspondence.
| Task | Who | Typical time |
|---|---|---|
| Read the request, identify the claims and date range | Billing lead | 1 hour |
| Export documents from the EHR, one at a time | Admin | 4–8 hours |
| Pull authorizations and UR letters from email and files | UR coordinator | 2 hours |
| Identify unsigned notes and chase signatures | Clinical director | 2–6 hours |
| Collate into order, build an index, number pages | Admin | 4–6 hours |
| Review for completeness before submission | Clinical director | 2–3 hours |
| Submit and log | Billing lead | 1 hour |
Call it sixteen to twenty-seven hours. Two to four working days, spread across four people, under a deadline that started before the letter arrived.
Nobody invoices for this, which is exactly why nobody manages it.
Why the hours are the smaller problem
Three things cost more than the time.
Displacement. The clinical director spent six hours chasing signatures. That is six hours not spent on supervision, on a utilization review, or on the admission that came in Wednesday. In a small program those hours have nowhere else to come from.
Discovery under deadline. The unsigned notes are found on day three, not day one. Some of them are two months old and the clinician has left. Now the response either goes out incomplete or goes out late, and both of those are findings.
Judgement under pressure. The completeness review happens on the last evening, by someone who has been at it for three days. That is not when anyone reads carefully.
The multiplier nobody plans for
The arithmetic above is for one patient. Requests frequently cover several, and a program that has attracted attention from one contractor often hears from others.
Two requests covering five patients each is not twice the work — it is the same four people, the same weeks, and a clinical director who is now doing this instead of their job for the better part of a month.
Next time a request arrives, ask everyone involved to note their hours against it. One line each, no timesheet. The number tends to surprise people, and it is the only version of this argument that will move a budget conversation.
What changes the arithmetic
Most of that list is not judgement work. Exporting, collating, indexing, and paginating are mechanical, and mechanical work is what software is for. The two steps that genuinely need a person — deciding what is in scope, and reviewing for completeness — are also the two steps that get squeezed when the mechanical work eats the week.
The version worth building toward is one where the compilation is instant and the signature gaps are known in advance, so the only remaining work is the part that actually requires someone's attention. That is not a marginal improvement in efficiency. It is the difference between a response assembled carefully and one assembled at midnight.