What RADV is

RADV stands for Risk Adjustment Data Validation. It is the program the Centers for Medicare and Medicaid Services (CMS) uses to verify that the diagnoses a Medicare Advantage (MA) plan submitted for risk adjustment payment are backed by real medical record documentation. In risk adjustment, each qualifying diagnosis maps to a Hierarchical Condition Category (HCC), and each HCC carries a coefficient that lifts the member's Risk Adjustment Factor (RAF) and the plan's payment.

A RADV audit works backward from payment. CMS samples member-year combinations from a contract, then requests the medical records that support every risk-adjusting HCC captured for those members in the audited year. If the record does not substantiate the diagnosis, the HCC is invalid and the associated payment is recouped.

Why RADV matters

RADV is the enforcement layer behind the entire risk adjustment economy. A capture program that lifts RAF on paper but cannot survive a records request is not revenue, it is contingent liability. The operational stakes are concrete: every HCC a plan books is a future audit obligation, and the documentation that justifies it has to be locatable years later.

The financial stakes changed sharply in 2023. The RADV final rule made audit findings extrapolated to the contract level and eliminated the fee-for-service adjuster. Extrapolation means a sampled error rate is projected across the full contract population rather than charged only on the sampled members. Removing the fee-for-service adjuster removed a buffer that previously offset part of the recovery. Together, these two changes raise the financial exposure for unsupported HCCs by orders of magnitude.

How RADV works

The audit cycle has a predictable shape, and each step is where exposure is created or contained:

  1. Sampling. CMS selects member-year combinations from a contract. Each sampled member carries a set of HCCs that the plan submitted for that payment year.
  2. Records request. For every sampled HCC, the plan must produce the supporting medical record from a qualifying face-to-face encounter in the relevant year, signed by an eligible provider.
  3. Validation. CMS reviewers check whether the record supports the diagnosis under coding rules. A captured HCC without retrievable supporting documentation is treated as unconfirmed and becomes a repayment risk.
  4. Extrapolation and recovery. Under the 2023 final rule, the confirmed error rate is extrapolated across the contract, and with no fee-for-service adjuster the recovery applies in full.

The defense against all of this is chain of custody per HCC: the source clinical note, the supporting labs or imaging where relevant, the provider attestation, and the coder review record, all attached to the captured diagnosis and retrievable on demand. Treated as a system requirement rather than a scramble at audit time, chain of custody is what turns a captured HCC into a defensible one.

Common mistakes teams make with RADV

  • Treating capture and defense as separate projects. Coding teams chase RAF lift while documentation retrieval is a downstream afterthought. By audit time the encounter, note, and attestation may no longer be locatable together.
  • Assuming a submitted diagnosis equals a supported one. A diagnosis that cleared the submission pipeline can still fail a records request if the underlying note does not substantiate it under coding rules.
  • Underestimating extrapolation. Teams still model RADV exposure as a per-member correction. Post-2023, a sample-level error rate applied across the whole contract is the real number.
  • Keeping evidence in disconnected systems. When the note lives in the EHR, the labs in a separate portal, and the attestation in a shared drive, no one can assemble a clean evidence packet per HCC quickly.
  • Auditing only after the year closes. Sampling documentation prospectively, while the encounter can still be corrected, is far cheaper than discovering gaps once the payment year is locked.

How Pelica handles RADV

Pelica's Risk Adjustment Copilot attaches a chain-of-custody evidence packet to every captured HCC as it is booked: the source note, supporting labs, provider attestation, and coder record, all on one canonical patient record rather than scattered across point systems. That means a RADV records request becomes a retrieval, not a reconstruction. Across Pelica deployments, customers have lifted RAF by roughly +0.4 in two quarters with no new headcount, and that lift is built to survive audit because the documentation is captured at the same moment as the diagnosis.

Related terms

RADV sits inside a tightly linked set of risk adjustment concepts. A captured diagnosis becomes an HCC, which feeds the member's RAF score and, through reconciliation cycles, surfaces in the MOR that confirms which diagnoses CMS actually recognized. Understanding how an HCC flows from capture to recognized payment is the same understanding that makes it RADV-defensible.

Sources