Medicare Advantage Substance Use Disorder Coding Fraud: When Coding Guidance Outruns the Record

Medicare Advantage substance use disorder coding fraud involves knowingly submitting unsupported substance-use diagnoses to increase risk-adjusted payments. Complete Health and Monogram Health settled federal allegations involving the same historical hierarchical condition category in 2026. For the clinician receiving the prompt, the question is concrete: what supported the diagnosis the coding program wanted?

This form of Medicare Advantage fraud reaches into the instructions a provider group gives its doctors. The condition-specific upcoding allegations concern the diagnosis requested, the record behind it and the payment it helped produce.

What HCC 55 and HCC 59 Meant in These Cases

The Complete Health allegations cover 2020 through 2023. The Justice Department identified HCC 55 as Drug and Alcohol Dependence and HCC 59 as Major Depressive, Bipolar, and Paranoid Disorders. Monogram’s alleged conduct ran from January 1, 2021, through December 31, 2023. Its announcement described HCC 55 as Substance Use Disorder.

Why the Diagnosis Needed More Than a Code

An eligible diagnosis feeds a hierarchical condition category, which contributes to the patient’s risk score. The score changes the payment Medicare makes to the plan. A provider’s contract can give the provider a share in that revenue. Taxpayers fund the inflated payment, while the patient carries a diagnosis the clinical record does not support.

The government alleged that diagnoses were clinically unjustified, lacked documentation or did not affect the patient’s care.

A coder may find the diagnosis in a record, while the physician’s encounter note raises a different question about why it belongs there. A coding instruction can explain which category to select while leaving the clinical basis unanswered. That is the gap the Complete Health allegations describe.

Complete Health’s Guidance Became Physician Prompts

Complete Health agreed to a $14,100,000 settlement announced on August 3, 2026, over alleged coding failures in those two diagnosis families. It is a management-services organization that manages, owns or operates health care provider groups in Florida, Alabama and Colorado. The alleged conduct therefore reached the way practices generated diagnosis codes, upstream of the plan’s payment submission.

In its Complete Health announcement, the Justice Department alleged that the company submitted codes that were not clinically valid, lacked proper support in the beneficiary’s medical records, or were not considered in the beneficiary’s care, management or treatment. The government described incorrect coding guidance followed by prompts asking doctors to add the diagnoses even when they were unsubstantiated or clinically unjustified. Doctors then added inaccurate codes, according to the allegations.

For someone teaching coding or reviewing physician queries, the wording of the instruction is central: what diagnosis did it ask the doctor to add, and what did the encounter support?

The whistleblower behind the case had worked as an associate director of risk adjustment at VIVA Health. A person working on the plan side could see a problem originating in provider coding. The information that exposes the mismatch can sit across the organizations sharing the payment.

Monogram’s Risk-Sharing Incentive Reached the Same Family

Monogram agreed to a $2,400,000 settlement announced on August 24, 2026, smaller than Complete Health’s despite overlapping substance-use coding allegations. The Justice Department’s Monogram announcement alleged that it knowingly submitted diagnoses that were not clinically accurate, lacked medical-record documentation, or did not require or affect patient care, treatment or management.

Alongside substance-use disorder, the government named HCC 21 for Protein-Calorie Malnutrition, HCC 48 for Coagulation Defects and Other Specified Hematological Disorders, and HCC 88 for Angina Pectoris. Substance-use coding was one part of a wider set of alleged documentation failures. A physician formerly employed by Monogram brought the whistleblower case and was to receive approximately 16% of the recovery.

The department expressly tied Monogram’s risk-sharing arrangements to a financial incentive to submit additional diagnoses. Higher patient risk scores meant higher payments from the Centers for Medicare & Medicaid Services. That is why the provider’s payment agreement belongs alongside the coding instructions when we examine the business behind the codes.

A 2015 economics paper, Upcoding or Selection? Evidence from Medicare on Squishy Risk Adjustment, hosted by Duke University, found that coding intensity increased with closer integration between insurers and providers. It describes how the payment incentive reaches the physician: an insurer can pay doctors according to the codes they assign instead of the visits and procedures they perform.

The paper also describes tools that prepopulate physician notes with prior diagnoses and training directed at billing staff. Its industry analysis explains why a coding program’s contracts, software and education materials belong in the same inquiry. Monogram’s settlement supplies the separate, company-specific allegation about the financial incentive.

The Behavioral Diagnosis Against the Encounter Record

The strongest comparisons put the submitted diagnosis beside what the clinician actually recorded. A current progress note can contradict a historical problem list. An audit can also find support for a depression diagnosis while rejecting the particular HCC submitted for payment.

A Problem List Overrode the Progress Note

A whistleblower brought the case involving Group Health Cooperative, Independent Health and DxID, and the government joined it. In its complaint, the government alleged that DxID coded, and Group Health submitted, major depressive disorder based on a problem list despite conflicting progress notes. The complaint set out what the physician had actually written: that the depression had resolved, had been a passing reaction to his sister’s death, and had eased enough earlier that year for him to ask to stop his antidepressant.

The allegation turns on a direct conflict within the patient’s records. According to the complaint, the coder selected the diagnosis from the problem list while the physician’s assessment described the depression as resolved. For a behavioral-health coding inquiry, that is a concrete comparison between the code selected and the clinical account available to the people selecting it.

The Depression Diagnosis the Anthem Audit Supported

A May 2021 audit of Anthem Community Insurance Company by the Health and Human Services Inspector General found that records did not support major depressive disorder for six sampled patient-years. In five of those cases, its independent medical reviewer found support for a lesser depression diagnosis that did not map to an HCC.

The audit’s comparison gives a behavioral-health insider a more precise question than whether depression appeared anywhere in the chart: what supported the particular diagnosis submitted for payment?

From the Requested Diagnosis to the Submitted Data

Connecting the instruction, the encounter note, any amendment and the submitted diagnosis is work we do with you. In a healthcare fraud case about coding, the final chart alone may hide how the diagnosis got there.

The Inspector General’s 2026 Medicare Advantage compliance guidance identifies physician prompts through electronic medical records, including prompts generated by artificial intelligence, among practices implicated in federal investigations. It describes prompts to add diagnoses patients did not have or that did not affect their care, treatment or management. The screen delivering the request is part of the evidence.

Amendment history can matter just as much as the final chart. In the separate Villages Health System settlement, the government described codes based on amendments that the treating provider had not initiated and that were untimely, or that the treating provider had not approved. That comparison adds questions about who changed the record and when to the question of what the diagnosis said.

The trail also includes what happened after someone identified an unsupported code. The Inspector General names failures to remove previously submitted invalid diagnoses, and reviews that add codes but leave errors standing recur in the enforcement cases. We connect the clinical objection to the later submission and payment decisions. A correction in the chart is only part of that account.

How the V28 Payment Model Changed Risk Adjustment

The 2024 CMS-HCC model, commonly called V28, incorporated clinical changes to HCCs as well as newer data. The 2026 payment announcement completed its phase-in for non-PACE organizations. Comparing older coding instructions with later submissions therefore requires the model and payment year for each.

The coding-intensity adjustment addresses differences in coding at the payment level. The patient-level inquiry still concerns the diagnosis, its support and what the organization did with it.

A whistleblower case under the False Claims Act can pay 15% to 30% of what the government recovers. These provider cases show why a clinician’s encounter knowledge and a risk-adjustment employee’s view of the submitted data each matter.

When the Prompt Asks for More Than the Visit Supports

If you know how a substance-use or psychiatric diagnosis moved from a coding instruction into a patient’s record, speak with our whistleblower experts about that instruction and what followed. We connect the clinical record, the coding decisions and the financial incentive to build your strongest Medicare Advantage whistleblower case. We stay with you through the whole case and bring in Medicare Advantage whistleblower lawyers who know provider coding programs.