Complete Health to pay $14.1M; MSO allegedly prompted doctors to add unsupported HCC diagnoses

On August 3, 2026, the Justice Department announced that Complete Health Partners Holdings, a management services organization based in Jacksonville, Florida, agreed to pay $14,100,000 to resolve False Claims Act allegations that it caused unsupported diagnosis codes to be submitted to the Medicare Advantage program. Most of the recent Medicare Advantage settlements named the insurers. This one lands on the coding operation itself, and the government's description of it reads like a coding-compliance case study.

What the government alleged

The government contends that Complete Health, a management services organization that operates affiliated provider groups in Florida, Alabama, and Colorado, was paid a percentage of what the Medicare Advantage plans collected from CMS. Because a higher risk score means a higher payment, that arrangement gave Complete Health a financial incentive to add diagnoses that raised its patients' scores.

From 2020 to 2023, the government alleges, Complete Health submitted diagnosis codes within two risk-adjustment categories that were "not clinically valid, not properly supported by the beneficiary's medical records, and/or not considered in the care, management, or treatment of the beneficiary." The two categories it named were HCC 55 (Drug and Alcohol Dependence) and HCC 59 (Major Depressive, Bipolar, and Paranoid Disorders).

Those are the government's allegations. DOJ states the claims resolved by the settlement are allegations only and there has been no determination of liability.

The coding pattern at the center of the case

The government laid the mechanics out step by step. It contends Complete Health "disseminated incorrect coding guidance to its coders and physicians" for the diagnoses in those two categories. Its coders then reviewed beneficiaries' records, flagged additional chronic conditions, and the company "prompted doctors to add those diagnosis codes, even when the diagnosis codes were unsubstantiated or not clinically justified." According to the government, the doctors added them, the codes weren't accurate, and the resulting risk scores drove higher CMS payments that the plans passed back to Complete Health.

Strip the legal framing and that's a workflow, not an accident. A coder surfacing a possible undocumented condition is normal risk-adjustment work. Handing a physician a diagnosis to sign that the record doesn't support is the line the alleged scheme is said to have crossed. That difference is the whole compliance question on a risk-adjustment desk.

Where a compliant query ends and this begins

A compliant provider query is allowed, and encouraged, to surface a gap. It presents the clinical indicators already in the record and asks the treating provider to confirm, clarify, or refute a diagnosis. What it can't do is lead. It can't supply the answer, tie the response to reimbursement, or push a specific code the evidence doesn't point to. The ACDIS and AHIMA compliant-query standard is built around exactly that boundary: a query is non-leading and evidence-based, or it isn't compliant.

The conduct the government describes runs the other way. It alleges the diagnosis was chosen first, by the coding operation, and the physician was prompted to add it. When the code comes before the clinical justification, a query has become a script, and every diagnosis it produces is an audit finding waiting to surface. Before any coder-suggested diagnosis goes on a claim, it needs independent, provider-authored, MEAT-level support in the encounter, not a coder's chart review with a signature attached.

The HCCs named, and where those diagnoses map under V28 today

One thing to keep straight: HCC 55 and HCC 59 are the numbers from the CMS-HCC V24 model, which was in effect during the 2020 to 2023 conduct. If you're coding today, you're on V28, fully operative since January 1, 2026, and V28 renumbered and split these categories. The conditions themselves still carry risk, so the lesson transfers directly. Here's how the diagnoses the case is about sit now (representative codes shown; the settlement named the HCC categories, not specific codes):

Two takeaways from the table. The single V24 substance-use category (HCC 55) is two separate HCCs under V28, split by drug versus alcohol, and both still adjust risk. The old omnibus mood-and-psychosis category (HCC 59) is likewise broken apart, with bipolar and major depression now scoring on their own lines. If you carried any of these forward from a pre-2024 chart, re-map them through the current model with the ICD-10-to-HCC tool before you rely on the old category number.

Why a risk-sharing arrangement raises the stakes

Recent Medicare Advantage settlements have mostly put the spotlight on the insurers. This one is about the entity a percentage of the premium flows to. The government's theory is that when a coding or management company is paid on a share of risk-adjusted revenue, every diagnosis it adds is also revenue to itself, and that incentive is what turned a coding operation into a False Claims Act defendant. If you work under a percentage-of-premium or shared-savings arrangement, the practical read is that a coder-suggested addend carries more audit exposure, not less, because the money motive is baked into the contract. Independent provider documentation is the only thing that neutralizes it.

How the case surfaced

The settlement resolves a lawsuit filed under the False Claims Act's qui tam whistleblower provisions, which let a private party sue on the government's behalf and share in any recovery. The relator, a former Associate Director of Risk Adjustment, will receive approximately $2,467,500 of the recovery. The matter was handled by the Civil Division's Commercial Litigation Branch, Fraud Section, and the U.S. Attorney's Office for the Middle District of Florida, with HHS-OIG. U.S. Attorney Gregory W. Kehoe said health care fraud enforcement "has long been a cornerstone of the mission of this office."

Where this lands on a coder's desk

This is a risk-adjustment case, so it touches your HCC mappings and RAF weights directly, unlike a straight medical-necessity matter. The self-audit is specific. Pull the diagnoses a coder or CDI reviewer added after the encounter, and for each one confirm the treating provider documented it independently with MEAT-level evidence, not a review note signed after the fact. Give the substance-use and mood-disorder categories a second look, since those are the exact families in this case and both still score under V28. And if a query in your shop ever names the diagnosis for the provider instead of presenting the evidence and asking, that's the pattern to fix now, before an audit letter makes the point for you.