Case Study 2 — Ninety-Eight and a Half: A Composite

Constructed. The practice, the acquirer, and the figures are not real. The situation — an organization reporting excellent collection metrics while losing money in three ways the metric cannot see — is ordinary, and it is the only case study in this book where nothing on any claim was wrong.**

Background

Section 23.10 posed the question this composite answers:

A practice reports a net collection rate of 98.5% — excellent by any benchmark. Name three serious problems that are entirely consistent with that number.

This is a practice that reported 98.5% for six years and then found out.

The composite

Constructed.

A twelve-physician specialty practice. Well managed, financially stable, and genuinely proud of its billing operation — with reason. Denials were worked promptly. Appeals were won at a good rate. Timely filing was never missed. Patient balances were collected.

Their net collection rate ran between 98% and 99%, year after year. It was on every board report. It was the number the practice manager was measured on, and she had earned it.

Then the practice was acquired, and a due diligence team ran a different set of reports.

What the due diligence found

Three things, none of which the net collection rate could have shown.

One — lost charges

The team compared the appointment schedule to accounts with charges, for a sample of dates.

(Constructed.) A meaningful percentage of kept appointments had no charge on the account.

Causes, all mundane: encounters completed after the billing cycle closed and never picked up; a hospital-based service performed by one physician that had never been interfaced to the professional billing system at all; and a category of small procedures that clinical staff believed were included in the visit and were not.

None of these ever appeared in a metric, because §23.9's rule holds: a charge that was never entered is absent from the numerator and the denominator both.

Two — underpayments written off as contractual

The team priced the practice's top thirty codes by volume against each payer's contract — §23.6's exercise — and compared the result to what had actually been paid.

(Constructed.) Two payers were paying below contract on a subset of codes, consistently, for years.

Every dollar of it had posted as a contractual adjustment. Which means it left the denominator, which means:

The underpayments were RAISING the net collection rate.

The worse the problem got, the better the number looked. Chapter 14's Case Study 2 was four years of exactly this, and Chapter 23 §23.10 states the mechanism.

Three — undercoding

A coding review of an encounter sample found a systematic downward shift on a share of E/M visits.

Chapter 15's Case Study 1 is the mechanism and this practice had a version of it: an internal audit worksheet that predated the 2021 rewrite, and coders scoring a framework that no longer governed.

The claims were coded correctly for what they said. They adjudicated correctly. They were paid in full.

Collection performance on an undercoded claim is flawless. The practice was simply billing for less than it had done.

Why the metric was not lying

This is the part worth being precise about, because it would be easy and wrong to conclude that the net collection rate is useless.

The net collection rate measures how well an organization converts entitled revenue into received revenue. Denials worked, appeals won, deadlines met, balances collected, accounts not carelessly written off.

This practice was genuinely excellent at that. The 98.5% was true and it was earned.

What the ratio cannot see is whether the entitlement was right.

Did we bill everything? (Charge capture — §23.9.) Did we bill it correctly? (Coding audit — Chapter 37.) Were we paid what the contract required? (Underpayment comparison — Chapter 28 §28.8.)

Three different questions, three different measurements, and none of them is a collection rate.

The failure was not that the practice reported the wrong number. It was that the number was reported as a report card for the whole revenue cycle, and the three questions it does not answer were therefore never asked.

What it cost

(Constructed.)

Recoverable: the underpayments inside timely filing windows, and the lost charges recent enough to bill. A fraction of the total, in both cases, and for the same reason Chapter 14's Case Study 2 and Chapter 18's Case Study 1 gave: these losses accrue continuously and their recovery windows are short.

Not recoverable: everything older. Six years of a metric that could not see them.

And a cost that is not a number: the practice manager had been measured on the one thing she was excellent at, and discovered at acquisition that the organization had been asking her for the wrong report for six years. That is not her failure, and she is the person most likely to experience it as one.

What it shows

First, this is the only case study in this book where nothing on any claim was wrong. Every prior failure was a claim that said something incorrect, incomplete, or unsupported. Here the claims were fine and the measurement was the failure.

Second, a single ratio became a report card, and that is the general lesson. No single number measures a revenue cycle, and the moment one is treated as though it does, everything it cannot see becomes invisible on purpose.

Third, the three blind spots are not obscure. Charge capture, coding accuracy, and contract compliance are the three most-discussed topics in revenue cycle management. The practice knew all three existed. It had simply never connected them to the number it was reporting.

Fourth — and this is the uncomfortable one — the metric moved in the pleasant direction as the problems worsened. Underpayments raise the net collection rate. The practice's best year on that metric may well have been one of its worst years in fact. Chapter 23 §23.10 and Chapter 15's Case Study 2 both say a favorable trend is a question; this is the sharpest version of it in the book.

And fifth, it was found by an outsider with different questions. Not a new employee, not a curious coder — a due diligence team whose job was to find out what the practice was actually worth, and which therefore asked questions nobody inside had a reason to ask.

That is the seventh finding in this book that came from a person rather than a control, and the second from outside the organization. The pattern is now too consistent to keep noting without answering, and Chapter 37 owes the answer.

The lesson

The net collection rate measures how well you collect what you billed. It says nothing about whether you billed everything, billed it correctly, or were paid what you were owed.

Four carry-forwards:

Report the net collection rate alongside three others: a charge capture reconciliation, a coding audit result, and an underpayment comparison. Four numbers, four different failures, and no single one of them can be a report card.

Reconcile the schedule to charges monthly. §23.9's first comparison. A signed note with no charge is one join and it is the query to build first.

Price your top codes against your contracts, annually. §23.6's exercise. It is an afternoon and it is the only thing that sees an underpayment written off as contractual.

And be suspicious of a metric that improves while nothing changes. Underpayments raise the net collection rate. If yours is rising and you cannot say why, that is the finding.

Discussion questions

1. The practice manager was excellent at the thing she was measured on. How should the organization have measured her instead? Be specific, and say whether your alternative is fair.

2. Underpayments raise the net collection rate. Name one other metric in health care — or outside it — that improves as the underlying situation worsens. What do such metrics have in common?

3. The three blind spots are the three most-discussed topics in revenue cycle management, and the practice never connected them to the number it reported. Why do you think that is?

4. Design the four-number dashboard. What are the numbers, how often is each produced, and what does each one cost to produce?

5. Seven findings in this book came from a person rather than a control, and two came from outside the organization entirely. State what you think the honest conclusion is — and say whether "build better controls" is a sufficient answer.