Case Study 1 — The Coverage Rule Nobody in the Laboratory Was Reading

A real, public matter — the Department of Justice's nationwide False Claims Act investigation into hospital billing for implantable cardioverter defibrillators furnished to Medicare beneficiaries, and the published resolution framework that ended it. Told from the documented record: a national coverage determination, a Justice Department review framework, and publicly announced settlements. This case study asserts no dollar figure and no participant count. Both are in the Department's own announcements, they were reported in waves over several years, and you should take them from the primary source rather than from a textbook — including this one.

Background

Cardiology is the specialty where §35.1's fourth item — the policy landscape — carries the most money per document, and this matter is the reason to believe it.

An implantable cardioverter defibrillator (ICD) is a device placed in patients at risk of sudden cardiac death from a life-threatening arrhythmia. It monitors the heart's rhythm and, when it detects one of the arrhythmias it is programmed for, delivers a shock. The clinical evidence supporting ICDs in defined populations is substantial, and the device is expensive: implantation is a hospital procedure with a device cost, a procedure, and an admission or outpatient encounter attached to it.

Because the population who benefit are defined by clinical trials rather than by intuition, Medicare covers ICDs through a national coverage determination (NCD) — the instrument Chapter 22 §22.3 taught: a nationwide, binding statement of what the program will and will not pay for, published by CMS and applied by every Medicare Administrative Contractor. The ICD NCD sets out the patient populations for which the device is covered, the diagnostic criteria that must be met, and — the part this case turns on — waiting periods.

Those waiting periods are the whole case. The NCD has, over its history, included provisions that a device implanted within a defined number of days after a myocardial infarction, or within a defined number of days after coronary artery bypass surgery or a percutaneous coronary intervention, is generally not covered — because the evidence supporting the device in those windows did not show the benefit that supported coverage elsewhere. A patient's heart function frequently improves during those weeks, which is precisely why the window exists: the device may not turn out to be indicated at all.

Read that as a coder. The procedure is real. The device is real. The physician's judgment is sincere. The documentation is complete and the diagnosis is accurate. The code is right. And the service is still not covered, because coverage is a rule about when, and the calendar is in a national coverage determination that lives outside the operating room, outside the chart, and outside the coding department.

The issue

Beginning around 2010, the Department of Justice opened a nationwide investigation into hospitals' billing of ICD implantations to Medicare, under the False Claims Act (31 U.S.C. §§ 3729–3733) — the statute Chapter 5 §5.3 built. The theory was straightforward and did not require anyone to have lied: claims were submitted for devices implanted inside the NCD's non-covered windows, and a claim submitted to a federal health care program is a certification that the service is covered and payable.

The scale of the inquiry was unusual. This was not a probe of one system; it reached hospitals of every size across the country, because the practice pattern was national and the underlying failure was structural rather than local. Hundreds of hospitals were eventually involved.

What happened

Two things about the Department's approach are worth a coder's attention, because they are documented and because they are not what people expect an enforcement action to look like.

First, the government published its criteria. In 2013 the Department released a medical review framework — widely referred to in the industry as the ICD "resolution model" — that set out how implantations would be categorized: which fact patterns it would pursue, which it would treat as falling within the NCD's own exceptions, and which it would not pursue at all. Hospitals could review their own records against the published categories and come forward.

That is a materially different posture from an audit. It converted a sprawling investigation into a self-review exercise with published rules, and it let organizations resolve matters on categories rather than on chart-by-chart argument. Chapter 37 §37.9's self-disclosure discipline and Chapter 37 §37.10's corrective action plan are the framework's descendants in a smaller organization's life.

Second, resolutions came in waves over several years, announced publicly by the Department, and they were settlements — resolving allegations without a determination of liability, which is how the overwhelming majority of False Claims Act matters end. The aggregate recovery was large. The number of participating hospitals was large. Go and read the Department's own announcements for both; this book will not print a figure it cannot stand behind, and an enforcement number quoted from memory is exactly the kind of assertion Chapter 5 warns you not to repeat.

And a genuine contest ran alongside it. Cardiology's professional community argued — publicly, and not unreasonably — that the NCD's waiting periods did not match every clinical situation a physician faces, that some patients inside a window are genuinely at risk in ways the rule does not anticipate, and that enforcing a coverage calendar against individual clinical judgments is a blunt instrument. That argument is not frivolous and this case study does not dismiss it. It is also, as Chapter 22 §22.1 established, an argument about whether the rule is right — not about whether it applied. The mechanism that exists for changing a coverage rule is reconsideration of the determination (Chapter 22 §22.4's only-mechanism-to-change-a-rule), and the mechanism that exists for a patient-specific exception is the one the NCD itself provides. Neither is a claim.

The outcome

The matter resolved. The NCD itself has been revisited by CMS since — coverage determinations are revised, and Chapter 22 §22.3 told you they are living documents rather than permanent ones. And the operational consequence in cardiology programs was immediate and durable: a coverage check moved into the scheduling workflow.

That is the whole fix, and it is almost embarrassingly small next to the size of the matter. Before an ICD is scheduled, somebody establishes the dates — the infarction, the bypass, the intervention — against the current NCD's windows, and documents the answer. It takes minutes. It happens before the patient is in a gown.

What it shows

The code was never wrong. Nothing in this matter was an ICD-10-CM or CPT error. The diagnoses were accurate, the procedures were performed and documented, and a coder reviewing the record would have found nothing to query. This is the purest example in the book of Chapter 22 §22.11's category — a claim that was never going to be covered, as distinct from a claim that was coded wrong — and it is why Chapter 29 §29.4's triage asks the coverage question before it asks the coding question.

A coverage rule that lives outside the department that generates the service will be missed. The NCD was public, free, binding, and unread by the people scheduling the cases — not through negligence but through org-chart geography. Nobody in a catheterization laboratory has a job description that includes reading national coverage determinations, and nobody in the coding department sees the case until after it has happened. The gap between those two facts is where several hundred hospitals' exposure lived. This is the "unowned seam" pattern the book has recorded since Chapter 17, and it is the largest instance of it.

Enforcement found this from data, not from charts. Implant dates and prior-event dates are both on Medicare claims. A calendar rule enforced against a claims database needs no chart review to generate a list — Chapter 21 §21.10's pattern detection at national scale, and the same structural fact Chapter 26 taught in miniature: an organization can be detectable from outside before it is detectable from inside.

And published criteria beat case-by-case argument for everyone involved. The 2013 framework let organizations sort their own records into categories, resolve the ones that fell inside the rule, and defend the ones that did not. That it worked is a genuine argument for something this book has said in smaller settings a dozen times: a written rule, applied consistently, is a defense, and its absence is not neutral.

Discussion questions

1. A coder reviewing one of these records before submission would have found nothing to query. State precisely what a coder could have caught, what they could not, and where in the organization the check belonged. Then name the front-end chapter that owns that kind of check and explain why its usual trigger — an authorization requirement — would not have fired here.

2. The professional community's objection was that a coverage calendar is a blunt instrument against individual clinical judgment. Take that argument seriously: what is the strongest version of it, what does Chapter 22 §22.1's distinction between clinical appropriateness and medical necessity do to it, and what is the legitimate mechanism for acting on it?

3. The Department published its review criteria before resolving the matter. Compare that with an ordinary payer audit as Chapter 37 describes one. What does publishing the criteria change for the organization under review, for the reviewer, and for the fairness of the outcome — and what would have to be true of a criteria set for publication to be safe for the government?

4. This case involved hundreds of hospitals with the same failure. Argue both sides: does the fact that the failure was universal make each individual organization more culpable or less? What does your answer imply about how a compliance program should treat "everybody does it this way"?

5. The fix was a dated coverage check moved into scheduling. Design it. Name who runs it, what three facts it establishes, what document it consults, where the answer is recorded, and what happens when the answer is "not covered" — including what the patient is told, by whom, and under which chapter's notice discipline.

6. Chapter 35 §35.10 tells a new specialty coder to pull the coverage policies for the specialty's top services in the first two weeks. Would that step have found this? Say honestly what it would have surfaced, what it would not have, and what an individual coder is actually able to do with a finding that belongs to a department they do not work in — naming the chapter that addresses being right and not being heard.