Case Study 2 — The Report That Got Received Well: Two Composites

Constructed. Both organizations, both people, and all figures are not real. The two paths — an internal report handled well and an internal report handled badly — are both ordinary, and which one an organization gets is decided almost entirely in the first conversation.

This is the only case study in this book that runs two versions of the same event.

The event, common to both versions

Constructed.

A coder at a surgical practice notices, while working a batch of claims, that a particular code pair carries modifier 59 on essentially every occurrence.

She checks. The NCCI file shows the pair, with a modifier indicator of 1. She pulls five operative notes. In three of them she cannot find anything establishing that the services were distinct.

She has done exactly what Chapter 21 §21.10 says to do: she has a pattern, not an instance, and she has enough to raise it and not enough to conclude anything.

She takes it to her manager.

Version A — the report that got received badly

Constructed.

Her manager listens, and then says three things, none of them malicious and all of them wrong:

"That's how we've always billed it." Chapter 15's Case Study 1 heard the same sentence about an audit worksheet, and it means the same thing here: the practice's history has been offered as an answer to a question about its correctness.

"The claims are paying." Chapter 14's Case Study 2 retired this one — a paid claim is not a verified claim — and by this point in the book it should produce an involuntary flinch.

"Let's not go looking for problems." This is the one that matters. It is not a refusal; it is worse. It tells the coder that the organization would prefer not to know, which is a sentence about the organization rather than about the claims.

What happens next

She stops raising things. Not dramatically — she simply does not bring the next one, or the one after. The organization has just disabled its most effective detector, and it will not notice for years, because a report that is not made leaves no trace.

The pattern continues. So does the exposure, and now with an additional feature: someone in the organization has been told about it. Chapter 5 §5.3's "reckless disregard" and "deliberate ignorance" are not abstractions here. An organization that has been told and declined to look is in a materially different position from one that never knew.

And the coder now has a decision she did not want. Section 21.10 mentioned that qui tam actions commonly originate with an employee who noticed a macro. They very rarely originate with an employee whose first report was welcomed. The whistleblower cases that reach the public are, overwhelmingly, the second or third step — and the first step was a conversation that went like this one.

Version B — the report that got received well

Constructed. Same coder, same pattern, same five notes.

Her manager says four things:

"Thank you — put it in writing and send it to me and to compliance today." Not because it is serious, but because the date matters and because a written report cannot be forgotten or misremembered.

"Don't pull any more charts." Chapter 21 §21.10's third rule. She has enough to raise it; anything further is an investigation, and that is not her job. This instruction protects her, protects the organization, and speeds everything up.

"What's the override rate?" The manager asks for the pattern rather than the instances — one query against claims data, no charts, Chapter 17's Case Study 1 technique. It comes back at 94%.

And: "We stop appending it automatically today, while we look." Not a conclusion. A pause, which costs the practice denials in the short term and stops the exposure from growing.

What happens next

Compliance reviews it with counsel's involvement, and reaches a conclusion — whatever that conclusion is. If claims were wrong, Chapter 5 §5.1's sixty-day clock is running from a date the organization can document, and the self-disclosure route Chapter 37 §37.9 lays out is available.

The macro comes out. Modifier 59 now requires a coder to open the note.

The surgeons get a one-page list of the sentences that are load-bearing in their common procedures — Chapter 17's Case Study 1 remedy, arriving before the audit rather than after.

And a monthly modifier frequency report starts running. One query.

Total cost: a week of several people's attention, plus whatever repayment is owed.

What separates the two versions

Not the facts. Identical.

Not the coder. Same person, same competence, same report.

Not even the manager's intentions — the Version A manager was not corrupt. He was busy, he did not want a problem, and he said something ordinary.

What separated them was about ninety seconds of the first conversation.

Three specific things the Version B manager did:

He treated a report as information rather than as an accusation. Nothing the coder said implied wrongdoing by anyone. The Version A manager heard it as a criticism of the practice, which is the single most common way a good report gets killed.

He asked for the measurement. A pattern claim has a number, the number is one query away, and asking for it converts an uncomfortable conversation into a factual one.

And he stopped the bleeding before deciding anything. Turning the macro off is reversible, cheap, and does not concede anything. The instinct to wait until you know before you change anything is exactly backwards when the thing is still happening.

What it shows

First, an organization's compliance posture is not its policies. Both organizations in this composite have a compliance program, a code of conduct, and a hotline. They differ in what happens when somebody uses the ordinary route first, and the ordinary route is what almost everybody uses.

Second, "let's not go looking for problems" is a legally distinctive sentence. Chapter 5 §5.3's standard reaches deliberate ignorance, and an instruction not to look is very close to the paradigm case. A manager saying it is exposing the organization more than the underlying claims were.

Third, the coder in both versions did exactly the right thing. She had a pattern, she had checked the public file, she had not over-investigated, and she raised it to her manager. Version A is not a story about her doing something wrong. It is a story about what a correct action runs into.

Fourth, the pause is the underrated move. Stopping an automatic override while you investigate costs denials and concedes nothing. Most organizations' instinct is to keep going until they are sure, which maximizes exposure during exactly the period they know they might be wrong.

And fifth — this book has now recorded five findings that came from a person rather than a control. Chapter 19's Case Study 2 said the absence of a place to say something is a control failure. This case study says the rest of it: having a place is not enough if the first person you say it to is annoyed.

The lesson

Whether an organization finds its own problems is decided by what happens in the first ninety seconds after somebody raises one.

Four carry-forwards:

If you are the person raising it: pattern not instance, in writing, promptly, keep a copy, do not investigate further. Section 21.10's three rules, and they are as much for your protection as for the organization's.

If you are the person receiving it: thank them, ask for it in writing, ask for the measurement, tell them to stop pulling charts, and pause the automatic behavior while you look. Four responses, ninety seconds, and they determine whether you ever hear from this person again.

Never say "let's not go looking for problems." Not because it is unkind. Because it is the sentence a plaintiff's lawyer would most like you to have said, and because it is usually the last thing an organization hears before it stops being told anything.

And measure whether people are raising things. An organization with zero internal reports does not have zero problems. Chapter 37 §37.8 makes this argument at length, and it is the single most reliable indicator of a compliance program's actual health.

Discussion questions

1. The Version A manager was not corrupt. Describe, precisely, what he did wrong — and say whether you think you would have done differently under time pressure.

2. The coder's report was identical in both versions. Is there anything she could have done to make Version A more likely to go well? Should the burden be on her?

3. "Let's not go looking for problems" is described as legally distinctive. Explain why, using Chapter 5 §5.3's language.

4. The Version B manager paused the automatic override before knowing anything. Argue against that decision, then say why the argument fails.

5. An organization with zero internal reports is described as a warning sign rather than a good sign. Is that fair? How would you distinguish an organization with no problems from one where nobody speaks?