Who Bears the Cost When Medicare Advantage Chart Reviews Inflate Risk Scores
Chart reviews let Medicare Advantage plans add diagnoses after the visit, and each added code can raise the risk score that sets federal payment. The question this piece settles is where that money actually goes: into plan margin, into reinsurance recoveries, or back out through audits that claw back a fraction of what was paid.
Diagnosis Codes That Pay Differently
In Medicare Advantage, the federal government pays a plan a benchmark amount adjusted by the risk score of each enrolled member. A sicker member generates a higher payment. Chart reviews exist to catch diagnoses the treating clinician documented but did not code on the claim. When a review adds a code, the member's risk score rises and the plan's payment follows.
The mechanics matter because the added diagnosis does not change the care the member received. A review vendor scans the medical record, finds a condition the encounter did not bill, and submits it. CMS pays the higher benchmark for the rest of the payment year. The enrollee's copay and network do not move in step.
Plans argue the documentation was incomplete at the point of care, and that the added codes reflect real conditions. Auditors counter that a code unsupported by the record is a payment error regardless of intent. The dispute is not about whether the patient is sick. It is about whether the chart supports the specific code submitted.
Hierarchical condition categories (HCCs) determine the risk score. Not every diagnosis maps to a paying HCC; only those that predict future cost do. A chart review that adds a nonspecific code may not move the score at all, while a code that maps to a major HCC can lift payment substantially. Review vendors prioritize conditions that carry HCC weight, and the difference between a paying and nonpaying code is the difference between a chart review that costs money and one that makes money.
Following the Risk-Score Premium Flow
Higher risk scores raise the benchmark payment the plan receives for each member. That money comes from the Medicare trust fund, not from the enrollee's monthly premium. So the incentive to add codes sits with the plan, while the cost sits with the federal budget. This is the core of the follow-the-money question.
Plans that run aggressive review programs tend to post healthier margins in their Medicare Advantage segments. The pattern shows up in segment reporting where medical loss ratios stay below peers. A lower medical loss ratio means more of the benchmark dollar is retained rather than paid out in claims.
Ceded reinsurance complicates the picture. A plan that cedes a block of business to a reinsurer shifts both the premium and the risk. If the risk score is later restated downward, the recoveries booked against that treaty can mask what the plan actually retained. A related piece on this site covers how AI underwriting changes disability premium ceded to reinsurers, where similar cession mechanics apply.
What Audits and Court Records Show
Office of Inspector General reports have repeatedly flagged Medicare Advantage overpayments tied to unsupported diagnoses. The pattern is consistent across audit cycles: reviewers find codes that the medical record does not support, and CMS issues a recovery. The recovered amount is typically far smaller than the estimated overpayment across the whole program.
Department of Justice settlements have cited unsupported diagnoses as a basis for False Claims Act liability. In these matters, the government alleges the plan knew or should have known the codes were unsupported. Settlements resolve the claim without a trial, so the underlying coding practices are described in the complaint rather than proven in court.
Whistleblower suits filed by former review staff describe production quotas for added diagnoses. Those allegations are unproven in most cases, but they sketch a workflow where reviewer pay depends on codes added. CMS Risk Adjustment Data Validation audits recover a portion of estimated overpayments, and the recovery rate is the number that matters for reserve modeling.
RADV audits sample a subset of members and extrapolate the error rate across the contract. The extrapolation is the contested part: a plan may concede errors on the sampled charts while disputing the projection to the full population. The method has been litigated, and the outcome shapes how much a plan reserves for a given audit cycle. A plan that wins on extrapolation can see its exposure drop by an order of magnitude relative to the government's initial demand.
Reinsurance Recoveries and Ceded Risk
Stop-loss treaties price off the risk score of the covered population. If the score is inflated, the treaty is priced on a population that looks sicker than it is, and the plan pays a premium that reflects that. When audits restate the score, the loss ratio the treaty was priced on no longer holds.
Recoveries from reinsurers can mask the true loss ratio in a reporting period. A plan that books a reinsurance recovery against an inflated claim base shows a lower net loss than the underlying experience supports. Retrocessional capacity depends on clean data, and a restatement forces the plan back to the market with a revised profile.
Audit adjustments create reserve surprises because the timing of a recovery is uncertain. A plan may hold a reserve for a known audit exposure, then find the final recovery exceeds it. The reverse happens too, and the variance lands in the quarter the adjustment is booked. This site has covered related reserve mechanics in Vermont's individual market rate filings and the All-Payer Waiver.
Taxpayers and Enrollees Bear the Residual
Overpayments reduce the Medicare trust fund, which is funded by payroll taxes and general revenue. The dollar that goes to an unsupported code is a dollar not available for covered benefits elsewhere. The enrollee does not see this directly, but the aggregate effect is a smaller pool for the program.
Enrollees face narrower networks and higher copays when plan margins compress. A plan that loses audit recoveries may respond by tightening prior authorization or trimming provider networks. The member experiences the adjustment as a change in access rather than a change in premium.
Providers absorb denied post-payment claims when a plan recoups a diagnosis it previously accepted. A physician group that coded a condition in good faith can face a take-back months later. The provider has no seat at the audit table and limited appeal rights in practice.
Practical Steps for Plan Oversight
Reconcile chart-review codes against the original encounter data before the risk score is submitted. A code that appears only in the review file and nowhere in the treating record is the one most likely to fail a RADV audit.
Track OIG findings and RADV recovery rates inside the reserve model, not as a separate compliance exercise. The recovery rate assumption drives the reserve, and it should be refreshed each audit cycle.
Stress-test stop-loss treaties for a risk-score restatement. Model the loss ratio the treaty would show if the score were reduced, and check whether the ceded premium still makes sense at the lower score.
Document provider attestation for every diagnosis added after the encounter. A signed attestation from the treating clinician is the strongest support for a code, and it is the first thing an auditor asks for.
Report suspected unsupported codes to the compliance function through a channel that does not route through the review vendor. The vendor that added the code has an incentive not to flag it.
Why Chart Reviews Persist
Chart reviews persist because the economics favor them. The cost to review a chart is a few dollars, while the payment increase from a single HCC can run into the thousands over a payment year. Even a low-yield review program can be profitable if the added codes survive audit. The calculation changes only when the expected recovery rate rises enough to erase the margin.
Plans also face a competitive pressure: if peers are adding codes and raising their benchmarks, a plan that abstains operates at a payment disadvantage. The dynamic resembles a race where the floor rises for everyone, but the trust fund pays for the increment. That collective action problem is why voluntary restraint has not taken hold.
Regulators have responded with audits and, more recently, with requirements for provider attestation. The attestation rule shifts the burden back to the clinician, who must confirm the diagnosis. In practice, attestation rates vary, and some plans report difficulty obtaining timely signatures. Where attestation is missing, the code is vulnerable on audit.
For investigators, the tell is a pattern of codes that appear only in review files and never in the treating record. That pattern, repeated across a contract, is what turns a documentation improvement effort into a fraud case. The line between the two is the support in the chart, not the intent of the reviewer.
This article is informational and does not constitute personalised insurance, legal, or financial advice. Readers should consult a qualified professional about their own circumstances.
