What the coding intensity adjustment is
The Medicare Advantage coding intensity adjustment is a reduction the Centers for Medicare and Medicaid Services (CMS) applies to every MA plan's risk scores. It corrects for a measured pattern: MA plans tend to code diagnoses more completely than fee-for-service (FFS) Medicare, so the same patient looks sicker on paper under MA. Left uncorrected, that coding difference would inflate risk scores and payment for a population that is not actually sicker.
Federal statute requires CMS to reduce MA risk scores by a minimum of 5.9% each year for this reason. CMS has applied that statutory minimum of 5.9% every year, including CY2026. The reduction is a flat payment haircut on risk scores, not a judgment about any individual plan or any individual diagnosis.
Why CMS applies it
CMS applies the adjustment because the same beneficiary tends to generate a higher risk score under Medicare Advantage than under fee-for-service. MA payment is built from risk scores, so a plan is paid more when its members carry more documented diagnoses. That gives MA plans a financial reason to capture every condition completely, and the data shows they do so more thoroughly than FFS Medicare, where the diagnosis on a claim does not change the payment for the visit.
The Commonwealth Fund describes the mechanism plainly: risk adjustment is meant to pay plans more for sicker members, but it also rewards more complete coding, so CMS reduces scores to offset the share of the difference that comes from coding rather than from real differences in health. Without the adjustment, the federal government would overpay MA plans for a population whose underlying health matches FFS.
The pattern is structural, not the result of any one plan behaving badly. FFS Medicare pays per service, so a provider has no payment reason to record every secondary diagnosis on a claim, and many true conditions go uncoded. MA pays on the risk score, so plans invest in chart review, annual visits, and coding programs that surface those same conditions. The result is two coding cultures producing different scores for the same underlying health, and the adjustment is CMS trying to neutralize that difference at the system level.
The 5.9% floor versus MedPAC's higher estimate
The 5.9% reduction is a statutory floor, not a measurement of how much extra coding actually happens. CMS has the authority to apply a larger reduction, but it has consistently used the 5.9% minimum, including in the CY2026 Rate Announcement. MedPAC, the independent commission that advises Congress on Medicare, has estimated that real coding intensity runs much higher.
In its March 2025 report to Congress, MedPAC measured MA coding intensity at roughly 16% above comparable FFS beneficiaries. The gap between the 5.9% CMS applies and the higher figure MedPAC measures is the center of the policy debate: MedPAC argues the floor leaves substantial overpayment in place, while CMS has held to the statutory minimum each year.
The reason the statutory floor and the measured figure diverge is partly design. Congress wrote the minimum into law, so CMS cannot go below it, and CMS has chosen not to go above it. MedPAC's estimate, by contrast, is an analytical measure of how much higher MA scores run than FFS for similar populations after accounting for differences in health. For a risk-bearing organization, the practical signal is that scrutiny of MA coding is rising, and the bar for documentation that holds up on audit is going up with it. An accurate score is worth more than a high one if the high one cannot be defended.
The two meanings of coding intensity
The phrase "coding intensity" gets used in two different ways, and they are easy to confuse. One is the CMS adjustment above, a flat payment reduction set in statute. The other is the general practice of how completely a provider organization documents and codes the chronic conditions a patient actually has.
- The CMS adjustment is a system-level payment correction. It applies across all MA plans regardless of how any single organization codes, and it is the same 5.9% for everyone.
- An organization's coding completeness is operational. It is how fully a medical group, HCC coding team, or risk-adjustment workflow captures the diagnoses a member genuinely carries, supported in the chart.
Accurate, complete, documented coding is legitimate and required under a risk-bearing contract. A condition a member has but no one documents and codes does not count toward the risk score, which understates the payment needed to care for that member and can leave a sick panel looking artificially healthy. The system-level adjustment and an organization's own completeness pull in opposite directions, and reading one as the other leads teams to either under-capture out of caution or over-reach for score. The problem to avoid is the second extreme: coding diagnoses that the record does not support.
Accurate capture versus over-coding, and where RADV fits
Accurate capture and abusive over-coding sit at opposite ends. Capturing the conditions a patient has, supported by documentation, is the job of risk adjustment. Inflating risk by coding unsupported diagnoses is what audits exist to catch. RADV, the CMS Risk Adjustment Data Validation audit, is the mechanism: CMS pulls a sample of charts and recoups payment for any HCC the documentation does not support.
The standard a coded condition has to meet is whether the note shows the condition was monitored, evaluated, assessed, or treated in the year it was coded. A diagnosis dropped onto a problem list without that support will not survive RADV. The practical guide to building capture that holds up is in building a RADV-defensible HCC chain. The goal is a defensible record for every code submitted, not the highest possible score.
How Pelica handles accurate capture
Pelica is the AI-native execution layer for value-based care: one live member record and a copilot beside every team that owns a risk-bearing contract. Most platforms show you what needs to happen. Pelica actually does it. The Risk Adjustment copilot reads the chart, maps each documented diagnosis to its HCC, and surfaces the conditions that are present in the data but not yet captured to a defensible standard. It works toward accurate, RADV-defensible capture, not inflated coding.
Because the copilot does the coordination and follow-up rather than handing a coder another worklist, capture happens inside the submission cycle. At HealthCare Partners, the largest IPA in the country, risk-adjustment coordination dropped from 30 minutes to 3 minutes per member, the same team now covers two to three times more members, and the organization added 0.4 to its RAF in two quarters with no new headcount, across 175,000+ patients managed live.
Related terms
For how a documented diagnosis becomes a member's risk score, see RAF and HCC. For the model that governs which diagnoses count and how completely they must be documented today, see CMS-HCC V28. For the audit that tests whether coded conditions hold up, see RADV.