A worked example: ninety days of aged AR, taken apart

Why this is an example and not a case study

We have no clients yet. A case study would therefore be invented client work, and an invented case study is the first thing a careful prospect checks and the last mistake a new company gets to make. So this is the honest version: a made-up aged AR report, labelled as made up, worked through exactly the way we would work a real one. You can judge the thinking, which is the only thing a case study was ever really for.

Every figure below is illustrative. None of it describes any practice.

The report

All of the following is invented for the purpose of showing a method. A four-provider pain management practice, $412,000 in total accounts receivable, $96,000 of it past 90 days across 340 claims — an average of $282 a claim in the bucket everybody is frightened of. Payer mix in that bucket: roughly three-fifths commercial, a quarter Medicare Advantage, the rest state Medicaid.

The first thing we do, which is not phoning anybody

We sort the 340 by recovery route rather than by age or by value, because the route decides who can fix it and how long it takes. On an illustrative mix like this the routes come out roughly: 90 claims denied for authorisation, 70 for documentation or medical necessity, 55 for modifier or bundling edits, 45 eligibility, 30 genuinely just slow, and 50 where the remittance says something that needs reading properly before anybody acts.

That sort alone changes the week. The 90 authorisation denials are one conversation with one payer repeated ninety times, and a person who has had it twice is four times faster at it than a person working alphabetically. The 30 slow ones need a statutory deadline quoted, not an appeal. The 50 unclear ones need ten minutes each with somebody who reads remittances, and half of them will turn out to be one of the other categories mislabelled.

What we would check before writing a single appeal

Two things, in this order. First, the filing limit on the Medicaid slice — a little over $17,000 of that bucket on this invented mix — because if the state allows 90 days rather than 365, some of it is already unrecoverable and the rest is urgent. Second, the statutory payment deadline in the practice’s own state for the commercial slice, because a claim past it is late against a statute rather than merely late, and that changes the phone call from a request to a citation.

Neither of those is work. They are two lookups, and skipping them is how a team spends three weeks appealing claims that died before they started.

The arithmetic we would put in front of the practice

At the published $57.23 per denied claim in administrative cost, working all 340 costs about $19,500 in labour. The published share of denied claims eventually paid when somebody works them is about 90%. Against $96,000 of billed value that is roughly $86,000 recoverable for $19,500 of effort — and those are somebody else’s measurements, cited on our calculator, not a promise about this imaginary practice.

The honest caveat we would say out loud on that call: the 90% figure is across all denied claims, not across claims already 90 days old, and age is the single biggest predictor of whether a claim ever pays. The real number for a bucket this old is lower and we do not know by how much. Anybody who tells you they do is guessing at your expense.

What we would not do

We would not start with the largest balances, because the largest balances are usually the most complex and the slowest to move, and a queue that shows no movement for three weeks loses the room. We would not chase the 40-odd balances under about $20 at all — at twelve minutes each they cost more than they return, and the hour belongs to the $900 surgical balance at 110 days. And we would not promise a recovery figure, because nobody can see a payer’s behaviour in advance and the vendors who promise it are pricing in the ones that fail.

The version of this that involves your actual numbers

Send an aged AR report and you get this analysis on your own figures, in writing, free, whether or not you buy anything. That is the whole offer and there is no second step — no sequence, no call unless you want one. It is also the only way either of us finds out whether the thinking above survives contact with a real book.