Case Study 1 — The Authorization That Didn't Match: A Composite

Constructed. The practice, the patient, and the figures are not real. The mechanism — an authorization obtained correctly, a service that drifted from it, and a denial nobody anticipated because everyone believed the matter was handled — is ordinary, and it is the denial category that produces the most surprise per dollar.

Background

Section 22.10 said three things about prior authorization, and this composite is the second and third of them:

An authorization is not a payment guarantee. It is tied to specifics — a code, a date range, a number of visits or units, a facility, sometimes a rendering provider — and a service that drifts from any of them may deny despite an authorization on file.

The composite

Constructed.

A patient is scheduled for an outpatient procedure. The practice does everything right.

Eligibility is verified. The medical policy is read. A prior authorization is requested with clinical documentation, and it is approved — for a specific procedure code, at a specific facility, within a specific sixty-day window.

The authorization number is in the chart, in the scheduling system, and on the claim.

What drifted

Four things, none of them a mistake by anyone.

The date. The patient rescheduled twice — once for a family reason and once because the facility moved the slot. The final date fell eleven days outside the authorized window.

The facility. The second reschedule moved the case to the practice's other site. Same practice, same physician, different facility, different identifier.

The code. Intraoperatively the physician found more than expected and performed a more extensive procedure. Clinically correct, correctly documented, correctly coded — and a different code from the authorized one.

And the rendering physician. The originally scheduled surgeon was unavailable on the new date, and a partner performed the case.

The claim went out with the authorization number on it.

What came back

A denial. (Constructed.) The remittance indicated the service was not authorized.

And the practice's response was the entirely natural one: "We have an authorization. Here is the number."

Which is true and irrelevant. The authorization was for a different code, at a different facility, by a different physician, on a date outside the window. The payer's system compared four fields and three of them did not match.

Why each drift is a different problem

And this is the part worth learning, because the four are not equally fixable.

The date. Frequently the easiest. Many payers will extend or reissue an authorization for a rescheduled service, especially with a documented reason — if asked before the service. Asked afterward, it becomes a retroactive authorization request, which some payers permit and many do not.

The facility. Depends entirely on how the authorization was issued. Some are site-specific and some are not. Nobody at the practice knew which this one was, because nobody had read that part of the approval.

The code. The hardest and the most defensible. A physician who finds more than expected and does more is doing medicine, not billing. Most payers have a process for this — a post-service review, or an authorization amendment — and the documentation supporting the intraoperative finding is the whole case.

And the physician. Frequently the easiest of all, because many authorizations are issued to a group rather than to an individual. Nobody checked.

Three of the four had a straightforward answer and nobody knew which questions to ask, because the approval document had been filed rather than read.

Where the patient came in

She had been told, before the procedure, that it was authorized.

That was true when it was said. It stopped being true across two reschedules that she participated in and neither she nor the practice recognized as a coverage event.

When the denial posted, the balance moved toward her, and Chapter 18's Case Study 2 established the mechanism: systems transfer balances when payers deny and appeals are exhausted, and nothing in the configuration asks whose fault the denial was.

She had rescheduled twice. (Constructed.) One of those reschedules was the facility's. She had no way to know that moving an appointment could void a coverage determination, because nobody tells patients that, because almost nobody in a practice thinks of it that way either.

How it was eventually resolved

(Constructed.) Partly, and slowly.

The practice requested a post-service review with the operative note supporting the more extensive procedure. That part succeeded — the clinical case was strong and the documentation was good.

The date and the facility were harder, and required escalation and a conversation with a provider relations representative.

Total elapsed time: several months. Total staff time: substantial. Total amount that would have been required to prevent it: one phone call at the time of the second reschedule.

What it shows

First, an authorization is a snapshot, and the world moves. It describes a service, on a date, at a place, by a person. Every one of those can change between approval and surgery, and three of the four changes in this composite were made by someone other than the patient or the surgeon.

Second, the reschedule is the event nobody treats as one. Practices have careful processes for obtaining authorizations and essentially none for re-examining them when something changes. The scheduling system and the authorization live in different places, and nothing connects a date change to a coverage record.

Third, the approval document should be read, not filed. Site-specific or not? Individual or group? Units or visits? Window length? Four questions, answerable in a minute, and the answers determine which drifts matter. Nobody at this practice could answer any of them.

Fourth, the code drift is the one that is right to accept. A surgeon who finds more and does more should do more. The system's job is to accommodate that, and most payers have a mechanism. A practice that pressures physicians to stay inside authorizations has solved a billing problem by creating a clinical one, and that trade is never worth making.

And fifth — the patient participated in the failure without any possibility of knowing it. This is the fourth case study in the book where that is true, after Chapter 16's observation stay, Chapter 18's maternity package, and Chapter 19's standing order. By now the pattern is not a coincidence and the book has said so; what it has not yet said is what to do about it. Chapter 26 owes that.

The lesson

A prior authorization describes a service, a date, a place, and a person. When any of them changes, the authorization is a question again.

Four carry-forwards:

Read the approval, not just the number. Site-specific? Individual or group? Window? Units? One minute, at the time of approval, and you know which changes will matter.

Build a trigger on reschedules. When a scheduled service with an authorization moves — date, site, or provider — somebody looks at the authorization. This is the whole fix and it is a workflow rule, not a system purchase.

Ask before the service, always. Nearly everything in this composite was solvable in advance and several parts were not solvable afterward. The asymmetry between "before" and "after" in prior authorization is larger than anywhere else in the revenue cycle.

And do not let this constrain the operation. The code drift was the physician doing their job. Accommodate it downstream.

Discussion questions

1. Four things drifted and three had straightforward answers nobody knew to ask about. Design the one-page checklist that would be completed when an authorization is received. What is on it?

2. The reschedule is described as "the event nobody treats as one." Where would you put the trigger, and who receives it? What is the false-positive cost?

3. The patient was told the service was authorized, and that was true when said. What should she have been told at the second reschedule? Write it.

4. The chapter says pressuring physicians to stay inside authorizations trades a billing problem for a clinical one. Is that too strong? Is there any version of that pressure that is legitimate?

5. This is the fourth case study in which a patient participated in a failure they could not have understood. The book keeps naming this and has not yet said what to do about it. What would you propose — and be specific about who pays for it.