By Marvin 
The $76 Billion Context
In its March 2026 report, MedPAC estimated that Medicare spends $76 billion more on Medicare Advantage enrollees than it would if those enrollees were in traditional fee-for-service Medicare. Of that overspend, $57 billion comes from favorable selection and $22 billion comes from coding intensity. Eight of the ten largest MA organizations had coding intensity scores five or more points above CMS’s adjustment factor. The overpayments increase Part B premiums for all Medicare beneficiaries by approximately $11 billion per year, roughly $175 per person.
These numbers matter because they represent the political and regulatory backdrop against which every retrospective coding program operates. Congress held dual hearings in January 2026 with insurer CEOs. The bipartisan UPCODE Act (S.1105) proposes excluding chart-review-only diagnoses and health risk assessment diagnoses from risk adjustment calculations, with CBO estimating $124 billion in savings over ten years. CMS finalized the exclusion of unlinked chart review diagnoses for CY 2027.
Retrospective chart review didn’t create this problem alone. But it’s at the center of the policy response, because add-only retrospective programs are the primary mechanism through which coding intensity accumulates year after year.
Why the Policy Response Targets Retrospective Specifically
The MedPAC data and Congressional action focus on retrospective coding because it produces a measurable, trackable signal. When a plan’s submitted diagnoses consistently increase through chart review activity without corresponding changes in clinical outcomes or service utilization, the data shows coding intensity rather than patient complexity. This is the signal that CMS monitors at the population level and that regulators cite when explaining enforcement actions.
Add-only programs amplify this signal with each review cycle. Every year, the program adds codes. Every year, the risk scores rise. The cumulative effect over multiple years is a coding profile that diverges from clinical reality by a widening margin. MedPAC’s finding that coding intensity accounts for $22 billion in excess payments quantifies the aggregate result of this pattern across the entire MA industry.
The policy tools targeting this pattern are already in motion. The unlinked chart review exclusion for CY 2027 means diagnoses found only through retrospective review and never linked to an encounter will no longer generate risk score credit. The UPCODE Act, if passed, would go further. The direction is unmistakable: retrospective programs that inflate codes without clinical grounding are losing their financial return as a matter of policy design.
What Survives the Policy Shift
Not all retrospective coding is under threat. CMS isn’t eliminating retrospective review. It’s eliminating the value of retrospective review that operates disconnected from clinical care. Diagnoses identified through chart review that are linked to legitimate encounters, supported by MEAT-based clinical evidence, and validated through two-way processes that add and remove codes still generate appropriate risk score credit. The model targets the how, not the what.
Programs that validate documentation quality before submission produce codes that survive both audits and policy changes. Programs that run two-way reviews produce coding profiles that reflect clinical reality rather than drifting upward on accumulated additions. Programs that connect chart review findings to provider workflows ensure that identified conditions are integrated into ongoing care, creating the encounter linkage that regulators require.
The difference between retrospective programs that survive the policy shift and those that don’t comes down to design. Programs designed for revenue accumulation lose value as policy closes the revenue pathway. Programs designed for accuracy and defensibility retain value because their output is what the model is designed to reward.
Restructuring Before the Window Closes
The CY 2027 unlinked chart review exclusion takes effect next year. The UPCODE Act is advancing through Congress. MedPAC’s $76 billion finding gives regulators and legislators the data they need to justify further restrictions. Plans running add-only retrospective programs that depend on unlinked diagnoses for revenue are watching their business model erode in real time.
Organizations restructuring around Retrospective Risk Adjustment Coding that links every diagnosis to an encounter, validates MEAT evidence before submission, and operates in both directions are building programs that retain value regardless of which policy changes come next. The codes they submit are encounter-linked, clinically evidenced, and defensible. That’s the standard that survives every regulatory scenario currently on the table, and the plans that meet it now won’t need to restructure again when the next restriction arrives.