Risk Adjustment Compliance Ignored: Audits Stopped, Refunds Abandoned and Records Left Uncorrected

Risk adjustment compliance ignored by management can leave unsupported diagnoses in the data used to set Medicare Advantage payments. You found the errors, explained their effect and put a correction in front of the people who could authorize it. Then the audit stopped, deletion work lost its priority or the approved fix changed a file the payment process never used. What management did with the finding is a revealing part of a risk adjustment fraud case.

The cases below concern decisions after someone raised the problem: which records could be reviewed, which findings would be acted on and which payments would be returned. For healthcare compliance, coding and finance staff, that sequence can connect the unsupported diagnosis to the company’s knowledge. A whistleblower case under the False Claims Act can pay 15% to 30% of what the government recovers.

The Correction Has to Reach the Payment

A risk-adjustment audit matters financially because the diagnosis helps set the plan’s monthly payment. Eligible diagnoses feed Hierarchical Condition Categories and the enrollee’s risk score. That score affects the amount Medicare pays the plan, and a provider’s contract can give it a share of the revenue. The clinical record, coding work and submission may pass through different organizations before the money arrives.

A review can identify a supported diagnosis that was missed, or an already-submitted diagnosis that the chart does not support. In one-way chart review, the additions move forward while the negative findings stall. Medicare pays for the additions while the organization keeps money the review says it should return.

Aetna agreed to pay $117.7 million in a March 2026 settlement. The Justice Department alleged that Aetna used chart-review findings to seek more money but ignored findings showing it had been overpaid. The settlement also covered morbid-obesity coding allegations brought by a former risk-adjustment coding auditor.

The HHS Office of Inspector General’s Medicare Advantage compliance guidance, issued in 2026, identifies failure to remove unsupported codes as a risk area. It describes responses to detected problems that include expanding the review, educating providers and submitting data for payment adjustments. It also recommends oversight before and after submission, including scrutiny of vendors whose compensation depends on diagnosis capture. The same guidance flags electronic-record prompts, including those generated by artificial intelligence, that push physicians to add diagnoses patients did not have or that did not affect their care.

The audit’s scope matters to what management learned. Which diagnoses, encounters and period did the reviewers examine? What did they count as an error? The next decision may be whether to examine similar records or earlier periods. A finding followed by a decision to stop looking tells a different story from a finding followed through to a correction and repayment.

Kaiser Had Already Shown That Correction Was Possible

A Kaiser Permanente whistleblower alleged that its Colorado operation corrected cancer coding and refunded payments after an internal audit, then stopped the correction. The allegation makes the abandoned control especially revealing: the organization had already demonstrated that it could find the problem and return the money. The whistleblower described a 2007 audit of more than 6,000 cancer claims, with reported error rates of 78 percent for breast cancer and 52 percent for prostate cancer.

A separate whistleblower allegation concerned a proposed question to the American Hospital Association’s Coding Clinic. According to that account, leadership decided not to send a question that would have resolved a disputed coding point.

In its January 2026 settlement announcement, the Justice Department alleged that Kaiser ignored physician concerns and internal compliance audits identifying inappropriate additions to medical records. It described aggressive diagnosis goals, attention directed at physicians who fell short, and bonuses tied to those goals. The alleged pressure reached the people treating patients and the facilities employing them.

Kaiser agreed to pay $556 million to resolve allegations involving diagnoses added after visits. The government alleged that many of those diagnoses had nothing to do with the visit to which they were added.

Sutter Turned the Review Back on the Reviewers

The Sutter Health whistleblower alleged that management answered a substantial audit finding by examining the auditors’ own work and stopping the audit. At one affiliate, the audit team found 383 errors across three conditions in a partial review of a single year. The compliance department then reviewed five of the team’s own encounters, found three errors and declared the team’s concerns unfounded.

The same complaint described a correction instruction that preserved the payment problem. According to the whistleblower, management told the team to change only the billing side of the electronic record. The medical record that the contracting insurer would use remained unchanged, and only the treating physician could correct the encounter. Management had been told that changing the billing file would leave inaccurate codes in the data going to the Centers for Medicare & Medicaid Services.

That is why the file named in a correction instruction matters. The audit team could finish the assigned task while the information used for Medicare payment stayed in place.

The government’s 2021 Sutter settlement announcement described its allegation more broadly: after learning of unsupported diagnoses, Sutter failed to take sufficient corrective action to identify and delete additional unsupported codes. Sutter agreed to pay $90 million. The provider received a share of the plans’ payments, putting the correction on its own revenue line as well as the insurer’s.

At Martin’s Point, Deletes Competed With Targets

The Martin’s Point whistleblower alleged that correction work lost to revenue goals even after the organization knew which codes were wrong. The former risk-adjustment operations manager described a manually assembled deletion list and a team worried that continuing to look for deletes would keep it from meeting targets. Many identified deletes allegedly never happened.

The complaint also described an internal review conducted in 2017 of charts from 2013 through 2015. According to the whistleblower, patients did not have, or their charts did not support, 60% of the illnesses reported to and paid by Medicare in that review. The whistleblower alleged that Martin’s Point neither broadened the sample nor searched earlier periods for the same errors. When the findings reached senior management that September, the whistleblower alleged, the discussion turned to generating more money rather than further review and correction.

The Justice Department’s settlement announcement addressed unsupported diagnoses from chart reviews submitted from 2016 through 2019. Martin’s Point agreed to pay $22,485,000 to resolve those allegations. The department said the relator, the former manager who brought the qui tam case, would receive approximately $3.8 million. Her job put her where the coding findings and the revenue targets met.

Group Health’s Warnings Reached the Decision Makers

The government’s complaint in the Group Health and Independent Health case describes adverse clinical findings followed by approval to submit the diagnoses anyway. Its account reaches beyond a disagreement between coders to the executives deciding what the plan would send.

The complaint alleged that at least 17 of 20 diagnosis codes captured by the chart-review vendor DxID and audited by Group Health failed documentation requirements. One example involved alcohol dependence drawn from an old medical-history entry that the record marked resolved.

In a separate review of old heart-attack diagnoses, the government described clinicians identifying patients who had not had the condition. It alleged that executives acknowledged concerns and nevertheless approved leaving the diagnoses in the submission. The warning had reached the people who could stop the codes going forward.

The same government complaint described an earlier correction at Independent Health that did not extend to later periods: after being told it could not code chronic kidney disease from laboratory values, the plan withdrew the codes for 2010 dates of service but left those for 2011 and 2012 in place, and it and DxID kept coding the condition from lab values for 2013 onward. The allegation concerns chronic kidney disease inferred from laboratory values alone. The government described notice, withdrawal for an earlier period and continued use afterward. The earlier withdrawal shows that Independent Health had already been warned about the coding it continued to submit.

Seoul Medical Acquired the Company That Proposed the Correction

The Seoul Medical Group whistleblower alleged that an outside management company’s audit led to a proposed withdrawal of codes, followed by an acquisition and an effort to prevent correction. Advanced Medical Management had calculated more than $6.7 million in overpayments for 2015 through mid-2017 and proposed a withdrawal timetable, according to the whistleblower.

The whistleblower alleged that Seoul Medical or its subsidiary bought the management company in June 2018 and then used personnel changes to prevent correction of its coding practices. Later that year, after the compliance committee raised the coding issue, Seoul Medical allegedly refused to allow review of charts prepared by its board members.

The government’s March 2025 settlement announcement described allegations involving spinal enthesopathy and sacroiliitis coded for patients who did not have those conditions. It also alleged that, when a plan questioned the spinal enthesopathy diagnoses, Seoul Medical enlisted Renaissance Imaging Medical Associates to create radiology reports that appeared to support them.

What Management Knew When It Stopped the Audit

The useful account of a stopped coding audit connects the finding, its recipients and the action that followed. An audit report shows what the reviewers examined and concluded. A meeting record or instruction can show who received that conclusion and whether they authorized a broader review, restricted the sample, changed the correction or ended the work. Whether you are a coder told not to delete unsupported diagnoses or a vendor reviewer who knows why additions earned money, we help connect what you saw to the decision that left the codes in place.

The submission history completes that account. It shows what reached Medicare after the audit, including whether the challenged code remained or a deletion was processed. A Medicare Advantage risk-adjustment data certification can bear the signature of the chief executive, chief financial officer or a delegate reporting directly to them. The audit findings available when that certification was made belong beside the data being certified.

Health care coding specialists have long distinguished finding an error from verifying a correction. In its discussion of hospital coding, Honest Mistake or Fraud? Meeting the Coding Compliance Challenge, published in the Journal of AHIMA in 1999, recommended retaining original, billed and post-billed code versions, together with the reasons for changes. That comparison shows what was recorded before review, what was sent for payment and what changed afterward. We work through that sequence with you, connecting what compliance, finance and submission staff each saw to the payment decision.

Why the Audit Can Become a Revenue Argument

Risk-adjusted payment gives an organization a financial interest in how thoroughly its patients’ conditions are coded. A study in Health Management Policy and Innovation compared Medicare Advantage and fee-for-service populations using both payment-based risk scores and another diagnosis-based measure that did not determine Medicare Advantage payment.

In Comparing Diagnostic Coding Intensity for Medicare Advantage and Fee for Service, published in 2019, the study examined claims from 2010 through 2014. It found that the alternative measure reduced the apparent risk difference between the populations. Its measured conclusion was that plans may be using more intensive coding to raise risk scores and, with them, payment rates.

The finding helps explain why a review’s financial direction matters. An addition can increase revenue. Removing an unsupported diagnosis can require money to go back. In the cases above, the decisive evidence concerns what happened when people inside the business identified that second outcome.

The Medicare Advantage fraud inquiry then becomes concrete: what did the organization learn, who could act on it and what did it continue to submit or keep? That is the account a compliance insider can bring to a whistleblower case.

When You Know Why the Correction Stopped

If you saw an audit narrowed, a deletion list abandoned or a refund stopped, speak with our whistleblower experts about the finding and the decision that followed. We work beside you to build your strongest Medicare Advantage whistleblower case around what management knew and what Medicare continued to pay. We bring in Medicare Advantage whistleblower lawyers who understand risk-adjustment coding and failed corrections, and stay with you through the case.