Case Study 1 — A Year of Unwinnable Appeals: A Composite

Constructed. The organization and the figures are not real. The failure — a billing office working a category of denial that could never be won, because nobody knew a field existed — is ordinary, and it is the only case study in this book where nothing was overbilled, nothing was underbilled, and the entire cost was effort.**

Background

Section 21.4 introduced the MUE Adjudication Indicator and said one thing about it plainly:

MAI 2 is the one people waste months on.

This is a place that wasted a year.

The composite

Constructed.

A hospital-affiliated multispecialty group. A capable denials team of six, with a work queue, a productivity standard, and a supervisor who tracked their appeal overturn rate.

Their overturn rate was good — except in one category.

Units-based denials. MUE denials overturned at a rate near zero, and had for as long as anyone tracked it.

The team's interpretation was that the payer was being unreasonable. They wrote better appeals. They attached more documentation. They escalated. One appeal for a units denial ran to eleven pages and included a letter from the performing physician.

Every one of them denied.

What was actually happening

A large share of the units denials carried an MAI of 2.

MAI 2 means the maximum is absolute — grounded in anatomy, in a code descriptor, or in a regulation. The units reported were not unusual. They were impossible, or forbidden as a matter of policy.

No documentation could change that, because documentation cannot make a patient have three of a paired organ, and it cannot make a code whose descriptor says "bilateral" describe four sides.

The team had been appealing facts.

Why nobody knew

The MAI is a column in a file nobody in the department had opened.

The denial did not say. The remittance carried a CARC and a RARC indicating the units exceeded the maximum. It did not carry the MAI, and neither the code nor the description distinguishes an absolute limit from a clinical benchmark. From the remittance, an MAI-2 denial and an MAI-3 denial look identical — and one of them is winnable.

The team's training had covered MUEs. It had covered them as "a maximum units limit," which is true and which is the half of the story that leads directly to writing appeals.

And the productivity standard rewarded working denials, not resolving them. Six people worked a queue. The queue got worked. Nothing in the measurement distinguished a denial that was resolved from a denial that was touched.

What it cost

(Constructed.)

Not a dollar of revenue. This is what makes the case study unusual. The claims were correctly adjudicated, the payments were correct, and no money was lost or wrongly received.

What was spent was a year of a denials team's time on a category that could not be won — time that came out of the categories that could. The queue is finite. Every hour spent on an MAI-2 appeal is an hour not spent on an underpayment, an indicator-1 override with real documentation behind it, or a timely-filing deadline.

Plus the eleven-page appeal, which required a physician's time. That one is worth naming separately: the billing office spent a clinician's attention on an argument that could not be made.

How it surfaced

A new analyst downloaded the MUE file.

She was building a report on denial categories and wanted to join denials to the codes' MUE values. The file had the MAI column in it, and she asked what it was.

Nobody knew.

She read the definitions, went back to a quarter of denials, and split them by MAI. The majority of the unwinnable category was MAI 2. The MAI-3 denials — a much smaller group — had actually been won at a respectable rate, which nobody had noticed because the categories had never been separated.

What it shows

First, this is the only failure in this book that cost no money and was still a serious failure. Every other case study involves an overpayment, an underpayment, or a patient harmed. This one is pure misallocated effort, which is invisible to every financial control an organization has, and which is probably more common than all the others combined.

Second, the remittance did not contain the information needed to work the denial. That is worth sitting with, because billing offices work from remittances. The CARC and RARC told the team a limit had been exceeded and gave them no way to know whether the limit was arguable. The field that decides it lives in a separate free file, and nothing connects them.

Third, "the payer is being unreasonable" is a hypothesis, and it was never tested. It is also the most comfortable available explanation for a denial you cannot win, which is precisely why it deserves suspicion. A year of zero overturns in one category is not evidence of an unreasonable payer. It is evidence that something about the category is not understood.

Fourth, the productivity measure was measuring the wrong thing. Denials worked is a volume measure. Denials resolved is an outcome measure, and the two diverge exactly where a category is unwinnable. Chapter 29 §29.7 covers denial metrics and this is the distinction it turns on.

And fifth — the finding came from a person again. A new analyst, building a different report, asking what a column meant. This is the fifth such finding in this book, and Chapter 19's Case Study 2 supplied the corollary that matters: it is not a control, but the absence of a place to ask "what is this column?" is a control failure. She asked because she was new enough not to assume everyone knew.

The lesson

Before you work a denial, find out whether it can be won. That question has an answer, in a free file, and it takes a minute.

Four carry-forwards:

Look up the MAI before touching an MUE denial. One field, three completely different correct responses. Section 21.4 is a page and it will save somebody a year.

Split your denial categories finely enough to see zero. An overturn rate averaged across a mixed category hides the fact that one sub-category is never won. The team's MAI-3 appeals were succeeding the entire time and nobody could see it.

Measure denials resolved, not denials worked. They are the same number until they are not, and the place they diverge is the place you most need to look.

And treat "the payer is unreasonable" as a hypothesis with a testable prediction. If it is true, some appeals in the category should succeed. If none ever do, the explanation is somewhere else.

Discussion questions

1. This failure cost no revenue. Is it therefore less serious than the overpayment in Chapter 20's Case Study 1? Argue a position and defend it against the obvious objection.

2. The remittance did not contain the information needed to work the denial. Whose problem is that? Should a remittance carry it? What would that cost the payer?

3. Design the denial-category taxonomy that would have made this visible. How fine is fine enough, and what is the cost of going finer?

4. "The payer is being unreasonable" was never tested. Write the test — what would you measure, and what result would falsify the hypothesis?

5. Five findings in this book have come from a person rather than a control, and this one came from somebody new enough not to assume everyone knew. What does that suggest about how long a new employee's usefulness as a detector lasts, and can anything extend it?