Keisha Wilson, CHC, CCS, CPC, CPCO, CRC, CPB, CPMA, AAPC-Approved Instructor
The message from the Office of Inspector General (OIG) is becoming increasingly clear:
Diagnosis coding is not simply about selecting a code. The medical record must support the diagnosis and accurately reflect the patient’s clinical picture.
Two Medicare Advantage compliance audits released by OIG on September 10, 2026, strongly remind us why these matters.
In its audit of HumanaChoice (Contract H5216), OIG found that medical records did not support the selected diagnosis codes for 178 of 220 sampled enrollee-years, resulting in $669,237 in identified overpayments. Based on its sample, OIG estimated that HumanaChoice received at least $130.9 million in overpayments for 2020 and 2021.
In a separate audit of UnitedHealthcare of Wisconsin (Contract H5253), OIG found that medical records did not support the selected diagnosis codes for 183 of 250 sampled enrollee-years, resulting in $722,280 in identified overpayments. OIG estimated at least $46.9 million in overpayments for 2020 and 2021 based on its sample results.
These are separate OIG estimates for separate organizations. Together, however, they illustrate a broader compliance issue: the diagnosis submitted for risk adjustment must be supported by the documentation.
And this isn’t an isolated area of OIG attention. OIG identifies a continuing series of targeted Medicare Advantage audits examining whether high-risk diagnosis codes submitted for risk adjustment are supported by medical records and comply with federal requirements. Three additional audits in this series are currently listed as active.
The Code Is Only as Strong as the Documentation Behind It
One of the most important concepts for healthcare professionals to understand is that a diagnosis appearing somewhere in the medical record does not automatically make it reportable.
The record should tell the clinical story.
What was the patient evaluated for?
What did the provider identify?
What conditions were assessed?
What was treated, monitored, evaluated, or addressed?
What clinical evidence supports the diagnosis?
What was the plan?
What services were actually performed?
And, where applicable, does the documentation support the medical necessity for those services?
CMS emphasizes that documentation should support the codes reported on the claim and that the medical record should include information such as the reason for the encounter, relevant clinical findings, assessment or diagnosis, rationale for diagnostic services, and plan of care.
That means we cannot look at a diagnosis code in isolation.
We have to look at the story behind the code.
Medical Necessity Is Part of the Conversation
Accurate diagnosis coding and medical necessity are closely connected, but they are not interchangeable concepts.
A diagnosis should not be added simply because it produces a more favorable reimbursement outcome.
Likewise, a service should not be billed at a particular level simply because the documentation contains enough words to make the claim appear to support it.
CMS identifies medical necessity as a primary criterion for payment and instructs providers to ensure that the medical record supports the services and codes reported.
This is why auditors should ask more than:
“Is the code in the record?”
The better questions are:
- Is the diagnosis clinically supported?
- Was the condition evaluated, treated, monitored, or otherwise addressed when required?
- Does the documentation support the specificity of the code?
- Does the service performed make clinical sense in the context of the patient’s condition?
- Does the documentation support medical necessity?
- Does the claim accurately represent what occurred?
That is where auditing becomes much more than a code-to-code comparison.
ICD-10-CM Guidelines Still Matter
The ICD-10-CM Official Guidelines for Coding and Reporting remain an essential part of the coder’s foundation.
The FY 2027 guidelines emphasize the importance of complete and consistent documentation and state that the entire record should be reviewed to determine the specific reason for the encounter and the conditions treated. They also emphasize that accurate coding cannot be achieved without appropriate documentation.
This is why coders should not rely solely on:
- Problem lists
- Templates
- Carry-forward diagnoses
- Historical diagnoses
- Auto-populated information
- Previous claims
- Risk-adjustment suspecting tools
- Diagnosis prompts
- A diagnosis appearing somewhere in the chart
Those items may help identify conditions for review, but they do not replace the need to evaluate the actual medical record and applicable coding requirements.
The record must support the code.
Coders, Auditors, and Billers Have an Important Role
This conversation is not limited to physicians.
Coders, auditors, billers, CDI professionals, compliance professionals, and other members of the revenue cycle team are part of the healthcare organization’s internal control structure.
Our responsibility is not simply to move claims through the system.
We have a professional responsibility to apply the applicable coding, billing, documentation, payer, and regulatory requirements to the work we perform.
That means speaking up when something doesn’t make sense.
It means asking questions when documentation does not support a diagnosis.
It means escalating concerns rather than simply changing a code without understanding why the discrepancy exists.
It means knowing when a diagnosis needs additional provider clarification and when the documentation simply does not support reporting the condition.
And it means understanding that compliance is not the responsibility of one department.
Fraud, Waste, Abuse and the False Claims Act
There is another reason this matters.
When federal healthcare programs are involved, inaccurate claims can have consequences beyond a routine denial or coding correction.
The False Claims Act provides liability for knowingly submitting, or causing the submission of, false claims to the government. The statutory definition of “knowingly” includes actual knowledge, deliberate ignorance of the truth or falsity, or reckless disregard of the truth or falsity; specific intent to defraud is not required.
That does not mean that every coding error is fraud or an FCA violation. A mistake, misunderstanding, or isolated documentation deficiency is not automatically evidence of fraud. But it does mean organizations need effective processes for identifying, investigating, correcting, and preventing inaccurate claims. And the current enforcement environment demonstrates why organizations should take these responsibilities seriously.
For example, in August 2026, the Department of Justice announced a $14.1 million FCA settlement alleging that a Medicare Advantage provider submitted false diagnosis codes to increase Medicare Advantage payments. Another August 2026 settlement involving Monogram Health addressed allegations concerning diagnosis codes that were not clinically accurate, not supported by documentation, and/or did not require or affect patient care, treatment, or management. These matters were resolved allegations and did not constitute adjudicated findings of liability.
Due Diligence Cannot Be an Afterthought
This is where organizations need to move from a reactive compliance model to a proactive one.
Don’t wait for:
- An OIG audit
- A payer audit
- A CMS request
- A denial trend
- An overpayment determination
- A whistleblower complaint
- A government investigation
before asking whether your diagnosis coding is supported.
Instead, build validation into the normal workflow.
Consider incorporating:
Prospective audits
Review documentation and coding before claims or risk-adjustment data are finalized when appropriate.
Retrospective audits
Identify patterns, vulnerabilities, and recurring documentation issues.
Provider education
Explain not only what needs to be documented, but why it matters.
Coder education
Keep coders current on ICD-10-CM guidelines, payer requirements, risk-adjustment requirements, and documentation expectations.
Bill review and reconciliation
Look for discrepancies between the medical record, coded diagnoses, services performed, and submitted claims.
High-risk diagnosis validation
Develop targeted review processes for diagnoses that present elevated compliance or reimbursement risk.
Corrective action and monitoring
When an issue is identified, don’t simply remove the code. Determine why it happened and whether similar claims may be affected.
The Goal Is Not to Code Less. The Goal Is to Code Accurately.
There is an important distinction here. Proactive auditing is not about telling providers or coders to undercode. It is not about avoiding appropriate diagnoses because they could increase reimbursement.
And it is not about creating fear around legitimate risk-adjustment coding. It is about ensuring the claim and the risk-adjustment data reflect the patient’s documented clinical condition and the services actually provided. If the patient has a condition, it should be appropriately documented and coded. If the condition is not supported, we should have the processes to identify that. If documentation needs clarification, there should be a compliant process for addressing it.
And if an organization discovers inaccurate data after submission, it should have a process for evaluating the issue and taking appropriate corrective action.
Audit. Educate. Validate. Correct.
The recent OIG findings involving HumanaChoice and UnitedHealthcare of Wisconsin are not simply statistics. They remind us that documentation, coding accuracy, clinical validity, medical necessity, and compliance are interconnected.
The medical record should tell the patient’s clinical story.
The codes should accurately represent that story.
The claim should reflect what was actually provided.
And organizations should have systems in place to identify problems before they become larger compliance and financial issues.
Audit readiness should not begin when the auditor arrives.
It should be part of the organization’s everyday culture. At KW Advanced Consulting LLC, we believe proactive auditing should do more than identify errors. It should help organizations understand why the issue occurred, educate the team, strengthen documentation and coding practices, and reduce the likelihood of recurrence.
Audit. Educate. Validate. Correct.
Resources



