Medicare Advantage Billing Corrections: Corrected in the Wrong File

A risk adjustment billing correction not sent to CMS can leave a Medicare Advantage overpayment intact. The billing team removes an unsupported diagnosis from its file, but the clinical record still carries it. If the plan takes its diagnosis data from that unchanged record, the correction stops before it reaches the payment. For the coder who found the error, that is a maddening result. The coder does the work. The wrong number survives. The Centers for Medicare & Medicaid Services can keep paying for a diagnosis the organization knows is unsupported.

A Sutter Health whistleblower alleged precisely that failure: auditors received instructions to correct the billing side while their authority to amend the clinical encounter was restricted. The distinction matters to anyone whose job ends at an internal correction but whose work feeds a health plan’s submission to Medicare.

Risk adjustment ties a Medicare Advantage plan’s payment to the enrollee’s diagnoses, and as the HHS Office of Inspector General explains, the risk score multiplies the plan’s base payment rate. The provider documenting the encounter and the insurer receiving Medicare’s payment can each have a financial interest in the same code while controlling different parts of its journey from diagnosis to revenue.

A chart review looks back through medical records for diagnoses relevant to payment, so a plan may get diagnosis information through both the original claim and a later review.

The correction question begins when review finds that an already-submitted diagnosis lacks support. In one-way chart review allegations, additions reach Medicare while findings that would reduce payment go nowhere. A billing-only correction can produce the same result through a different route: someone acts on the finding, but changes a file the submission process does not use.

Consider a simplified example. A visit produces a clinical encounter record and a billing record carrying the same diagnosis. An auditor identifies an unsupported code, and the billing team removes it from its record. The clinical encounter remains unchanged. A later extraction reads that clinical record, and the plan sends the diagnosis to Medicare. The billing correction and the submission now tell different stories about the same visit.

Sutter’s Auditors Were Told Which Side to Correct

At a Sutter affiliate, a whistleblower alleged that an audit team found 383 errors across three conditions during a partial review of a single year. The whistleblower said management instructed the team, on the afternoon before Thanksgiving in 2014, to make changes only on the billing side of EPIC. Only physicians could correct the patient encounter.

The alleged restriction put the person who found the error and the person allowed to correct the clinical record in different seats. The billing team could act on its own file. The clinical record the insurer later swept for diagnosis data could remain untouched.

The whistleblower also alleged that management had been told a billing-only change would not fix inaccurate codes in the data Medicare received. That acknowledgment gives the instruction its significance. The complaint describes an organization told why its chosen correction would stop short of Medicare.

Sutter later agreed to a $90 million settlement in August 2021. The Justice Department’s announcement described allegations that Sutter submitted unsupported diagnoses and, after learning of the problem, failed to take sufficient corrective action to identify and delete additional unsupported codes. A former employee brought the qui tam case, and the government joined the claims involving Palo Alto Medical Foundation.

A Corrected Billing File Can Leave the Clinical Record Untouched

In the Sutter whistleblower’s account, restricting clinical changes to physicians left auditors unable to finish the correction when a physician declined to amend the encounter or had left.

The distinction is visible in the records themselves. The billing edit shows what the billing team changed. A clinical amendment shows what happened to the encounter. The insurer’s extracted data show which version traveled onward. Counting completed billing corrections answers only the first of those questions.

There is also a difference between fixing the next extraction and correcting a diagnosis already sent. An amended source record addresses what a later review will find. The submission history addresses what Medicare has already received. Both belong in the same examination, because the date of an internal edit does not tell us what happened to an earlier payment input.

Who received the finding and who restricted the correction can connect this data problem to management’s response to an audit. The technical question remains specific: after the correction, which diagnosis reached Medicare?

Anthem’s Missing Comparison Left Earlier Codes Standing

The government’s complaint against Anthem alleged a different break in the correction path. The complaint says Anthem did not compare diagnoses it had already sent to Medicare with Medi-Connect’s chart-review results for the same patients and visits. The comparison could have identified previously submitted codes the review did not support.

The complaint described a comparison Anthem had the capacity to run:

The government illustrated the consequence with a patient whose active lung-cancer diagnosis the chart review did not substantiate. Anthem allegedly left that code in place while using the review to submit three additional diagnoses from the same visit. According to the complaint, Medicare then used the lung-cancer category in calculating payment.

The complaint describes review results relevant to both additions and corrections, but only the additions made it through. Where Sutter’s whistleblower described a correction confined to the billing file, the Anthem complaint described review results never reconciled with the prior submission.

For a failed correction, the evidence needed to report Medicare fraud connects the unsupported diagnosis to the audit finding, the internal edit and the data sent to Medicare.

We trace the same patient, encounter and code with billing, coding and data staff:

• The original encounter and audit finding show the diagnosis, the documentation reviewed and why the reviewer challenged it.
• The billing edit and clinical amendment history show which record changed, when it changed and who had authority to make the change.
• The plan or vendor extraction and submission history show whether the disputed diagnosis traveled onward or had already reached Medicare.
• The correction or deletion response shows what happened after the finding reached the submitting organization.

Red Flags Around a Failed Correction

A dated audit finding followed by a billing-only instruction and unchanged submission data is one of the red flags for Medicare fraud worth examining. An email acknowledging that the chosen edit will leave the submitted diagnosis unchanged can explain more than a tally of completed corrections. A billing specialist may know exactly what was changed. A data specialist may know exactly what the extraction reads. Together, those accounts can explain why the correction never reached the money.

A health plan can request the clinical record to support or validate the diagnosis data used for payment, and the risk-adjustment data regulation ties each submission to the provider’s records, the basis for a plan’s record requests and chart reviews.

For outpatient diagnoses, the Justice Department’s description of the documentation requirement includes whether the condition required or affected care, treatment or management at the visit. A correction inquiry needs that clinical support and the submission history together. The question is what Medicare was told about that encounter after the organization learned something was wrong.

The Incentive Follows the Diagnosis

Coding practices can change a risk score even when the patient stays the same. The 2015 working paper Upcoding or Selection? Evidence from Medicare on Squishy Risk Adjustment, hosted by Duke University, examined coding differences between private Medicare plans and fee-for-service Medicare. Its findings connected coding intensity to the relationship between insurers and physicians:

The organization producing the diagnosis and the organization collecting the payment can have aligned incentives. For an insider examining a failed correction, the useful comparison is how the business moves additions through that relationship and how it moves deletions.

Medicare Advantage cases have addressed unsupported diagnosis submissions and failures to correct them after the organization learned of the problem, and Sutter’s settlement shows a provider’s part in that chain under the False Claims Act.

The case turns on the connection between the unsupported submission, what the organization knew and what it did with that knowledge. An instruction limiting the correction, an acknowledgment of its effect and the later submission can tell that story together. We examine those decisions with you and bring in Medicare Advantage fraud whistleblower lawyers for the legal questions.

A whistleblower case under the False Claims Act can pay 15% to 30% of what the government recovers.

When You Know Where the Correction Stopped

If you corrected the billing file and know the diagnosis still reached Medicare, speak with our whistleblower experts about the correction that stopped short. We connect your knowledge of the records and instructions to the payment trail and build the case with you. We bring in Medicare Advantage fraud whistleblower lawyers who understand diagnosis submissions, and stay at your side as the case moves forward.