Medicare Advantage Coding Intensity Adjustment: A Lower Payment Does Not Fix the Diagnosis

The Medicare Advantage coding intensity adjustment reduces risk scores for differences in diagnosis reporting. It does not make an unsupported diagnosis supported. The Centers for Medicare & Medicaid Services still requires plans to submit accurate diagnoses and correct inaccurate data they learn about. If you work in coding or compliance and hear that the agency has already adjusted for the problem, the missing question is what happened to the diagnosis. A lower payment does not supply the clinical support that was missing when the plan submitted it.

That distinction matters in risk-adjustment fraud: a national payment correction and a decision to retain an unsupported code answer different questions. The agency makes that distinction itself, in the same announcement that says its adjustment is sufficient.

What Is Coding Intensity in Medicare Advantage?

Coding intensity in Medicare Advantage measures how diagnosis reporting differs from reporting for comparable patients in traditional fee-for-service Medicare. A patient can generate a higher risk score because a plan records more conditions, even without a corresponding difference in the patient’s health.

The financial connection is direct. A plan receives monthly payments adjusted for its enrollees’ expected costs. Eligible diagnoses enter hierarchical condition categories, whose payment factors contribute to the risk score alongside demographic and other characteristics. The plan’s base payment is multiplied by that score. The path from diagnosis to revenue can extend to a provider group sharing in the plan’s payments and a vendor paid to find additional diagnoses.

This gives the plan a reason to invest in finding conditions that ordinary billing may miss. The Medicare Payment Advisory Commission describes several causes of higher coding intensity, including more comprehensive documentation, fraudulent diagnosis data and other factors. The measurement combines those causes. For an insider, the useful question is which activity produced the additional diagnosis and what supported it.

An assessment, a physician’s visit and a retrospective chart review can all feed the payment process. Their labels tell us where information came from. The medical record tells us what the clinician documented, and the submission history tells us what the plan asked Medicare to pay for.

Medicare Advantage Coding Intensity Adjustment: CMS and MedPAC Disagree

The Centers for Medicare & Medicaid Services considers its uniform adjustment sufficient for calendar year 2026. The Medicare Payment Advisory Commission projects that a substantial coding-related payment difference remains in that same year. Taxpayers help pay for the gap the commission says remains.

The law has required a cut of at least 5.9% since 2019. The agency applied that minimum for 2026 and, in its rate announcement, said its analysis found the uniform reduction sufficient to reflect coding differences between Medicare Advantage plans and fee-for-service health care providers.

The Medicare Payment Advisory Commission reached a different conclusion. In its March 2026 report to Congress, it projected Medicare Advantage risk scores about 10% above those for comparable fee-for-service beneficiaries. After CMS’s adjustment, MedPAC projected a remaining 4% increase in risk scores attributable to coding intensity, producing $22 billion in higher payments for 2026.

That is MedPAC’s estimate of the remaining payment effect, with CMS’s reduction already accounted for. It measures the coding difference across populations.

The commission also recommends using two years of diagnostic data, excluding diagnoses documented only on health risk assessments, and applying an adjustment that closes the remaining coding gap. Those recommendations address both the information entering the model and the reduction applied afterward.

Favorable selection is another part of the payment debate: the people enrolling in a plan can be healthier than the payment formula captures. Coding intensity concerns how their conditions are recorded. Separating who enrolled from how they were coded matters before assigning a payment difference to either cause.

Why the Same Adjustment Lands Differently Across Plans

A uniform reduction reaches plans with different coding practices. A national average can therefore conceal the organizations where the incentive to add diagnoses has the greatest effect.

The 2015 working paper Upcoding or Selection? Evidence from Medicare on Squishy Risk Adjustment, hosted by Duke University, examined how to separate changes in coding from differences in the people choosing plans. It estimated that enrollees in private Medicare plans generated diagnosis-based risk scores 6% to 16% higher than those same patients would generate in fee-for-service Medicare. It also found greater coding intensity where insurers and providers were more closely integrated.

The paper explains one way the insurer’s payment incentive can reach the clinician:

The organizational connection matters: the plan receiving the risk-adjusted payment may have considerable influence over the people producing its diagnoses. Provider contracts, coding instructions and review incentives help explain how that influence operates inside a particular business.

A single adjustment does not measure those arrangements individually. The paper’s findings show why a correction based on the average can leave different plans in different positions. For a whistleblower, the company’s own handling of diagnosis findings is more informative than its place in an industry average.

What Did the V28 Transition Change?

The V28 transition changed the model used to calculate risk scores. It updated the underlying data and clinical groupings used to predict costs, including changes needed to build the model using ICD-10 diagnoses.

The Centers for Medicare & Medicaid Services phased in the 2024 CMS-HCC model, commonly called V28, over three years beginning in 2024. For the standard Medicare Advantage model, the transition completed in 2026, with the score calculated entirely under the new model.

MedPAC measured a substantial difference when it compared both models using the same year’s data. It calculated that coding intensity under V28 was 8.8 percentage points lower than under V24, the previous model, in 2024. That comparison shows how much the model itself can change the measured coding effect.

For a coding team, the practical point is that a diagnosis’s payment significance depends on the model and payment year. Changing the model does not change what a clinician documented in an earlier encounter or what the plan did after a review challenged that diagnosis.

How Does Normalization Differ From the Coding Adjustment?

Normalization accounts for changes over time in fee-for-service risk scores. The coding intensity adjustment addresses differences between Medicare Advantage and fee-for-service coding. The agency explains that the fee-for-service trend reflects changes in health, demographics and coding practices.

A finance team may see normalization, the coding adjustment and a model update together in its revenue forecast. Each changes the calculation for a different reason. None supplies a missing visit note or resolves a clinician’s objection to a diagnosis. Those questions remain attached to the patient record.

Health Risk Assessments Put the Record in View

Health risk assessments and chart reviews supply diagnoses beyond those captured through ordinary encounter reporting. An assessment may take place in the enrollee’s home through a plan employee or vendor. A chart review looks backward through medical records for diagnoses that were not included in the original billing.

The Department of Health and Human Services’ Office of Inspector General estimated in its October 2024 report on health risk assessments that diagnoses appearing only on assessments or assessment-linked chart reviews, and on no other 2022 service records, generated $7.5 billion in risk-adjusted payments for 2023. The Inspector General identified two possible explanations for the missing corroborating services: inaccurate diagnoses producing improper payments, or people with serious conditions who did not receive needed health care.

That distinction puts the patient back into the payment discussion. A condition recorded to raise payment raises a further question about the care that followed. The assessment, its clinical support and the subsequent care records matter together.

Twenty companies accounted for 80% of the total. The concentration makes the operation inside a company worth understanding: who performed the assessment, what the form documented, and how its findings reached both the payment process and the people responsible for care.

A prospective change in accepted data sources addresses future calculations, and the documentation and correction history still answer what happened to an earlier submission, whatever the 2027 rules now say about unlinked chart review records. For an earlier diagnosis, the patient record still has to supply the clinical support.

What a National Adjustment Cannot Tell You About a Code

A national adjustment cannot tell you whether a particular review found an unsupported diagnosis, who received that finding or why the code stayed in the payment data. Those are decisions inside the business.

The government’s complaint against Anthem alleged that its one-way chart review sought additional revenue while disregarding findings that could reduce payments. The complaint says Anthem told its vendor, Medi-Connect, to concentrate its chart review and coding on finding new codes that would bring in revenue.

The government also alleged that Anthem had no process to compare previously submitted diagnoses with the vendor’s results for the same patients and visits, although its programmers could have built that comparison. The alleged omission concerned what the review was designed to find and what the company chose to do with its results.

A revenue forecast may already include CMS’s adjustment. The review instructions can still direct employees toward additions while leaving unsupported diagnoses untouched. We examine the clinical record, the applicable model, the submission and the response to the finding together.

The Centers for Medicare & Medicaid Services uses Risk Adjustment Data Validation audits to examine improper payments. The audit and payment-validation process addresses a different question from setting a national coding adjustment. Inside a company, compliance findings and decisions about correction can show what the organization learned before any government audit reached it.

CMS Still Requires Plans to Correct Inaccurate Diagnoses

The agency expressly says the coding adjustment leaves plans’ accuracy obligations in place. Its 2026 announcement requires diagnoses supported in the medical record and coded under the applicable guidelines, including diagnoses from assessments and chart reviews. It also says an organization receiving information about inaccurate diagnoses must correct its data.

That is why the Medicare Advantage fraud inquiry returns to the submission and the response to contrary evidence. The payment model can change, and the national adjustment can remain contested, while the company has its own unresolved finding about a particular diagnosis. We work with insiders to connect that finding to the payment and the decision about correction. A whistleblower case under the False Claims Act can pay 15% to 30% of what the government recovers.

When the Adjustment Becomes the Answer to Every Objection

If your healthcare organization invokes the coding adjustment while leaving unsupported diagnoses in its submissions, speak with our whistleblower experts about the findings it set aside. We work beside you to build your strongest Medicare Advantage whistleblower case around the diagnoses, the money and the decisions you know firsthand, and bring in Medicare Advantage whistleblower lawyers who understand risk-adjustment coding and the V28 transition.