The Codes That Age Out
Every retrospective chart review produces codes from historical clinical documentation. A chart from January 2024 generates HCCs based on what was documented during encounters in that calendar year. Those codes get submitted to CMS. The following year, another retrospective review cycle begins. The question nobody asks often enough: are the codes submitted last year still valid?
Clinical conditions change. A patient with CKD Stage 3 in 2024 may have progressed to Stage 4 or stabilized with treatment. A patient diagnosed with major depression may have achieved remission. A patient whose morbid obesity was documented at a specific encounter may have lost significant weight. The code submitted last year reflected that year’s clinical reality. But if the plan carries that code forward into the next submission cycle without verifying that the condition is still active and managed, the code becomes stale, and stale codes fail audits.
The OIG’s March 2026 audits found that the most common failure pattern was history-of conditions coded as active diagnoses. That pattern is a direct product of carrier-forward coding: a diagnosis enters the system during one review cycle and persists in subsequent cycles without fresh validation.
How Carrier-Forward Coding Creates Population-Level Drift
The carrier-forward problem compounds at scale. Each year, a retrospective program adds new codes from current-year charts. But prior-year codes remain in the submission set without systematic re-evaluation. Over three or four years, a member can accumulate HCCs that no longer reflect their actual clinical status. The plan’s risk score for that member drifts upward not because the patient got sicker, but because old codes never got removed.
Multiply this across thousands of members, and the plan’s overall risk profile diverges from clinical reality. CMS monitors this divergence. When a plan’s risk scores climb without corresponding changes in hospitalizations, emergency department visits, or treatment costs, the data tells a story of coding activity rather than clinical complexity. That’s the population-level signal that triggers audit selection.
The Aetna DOJ settlement ($117.7 million, March 2026) covered coding practices spanning 2015 through 2023. That’s eight years of accumulated codes. The longer add-only programs run without systematic re-validation, the wider the gap between submitted codes and clinical reality, and the larger the financial exposure when enforcement catches up.
Breaking the Carrier-Forward Cycle
The fix is annual re-validation. Every code carried forward from prior submissions should be evaluated against current clinical documentation before being re-submitted. Does the member still have an active diagnosis with MEAT evidence in this year’s medical records? Is the condition being actively managed? Has the clinical status changed in a way that affects the HCC mapping?
AI tools that evaluate current-year documentation against prior-year submissions make this practical at scale. The system compares this year’s charts against last year’s submitted codes and flags discrepancies. Conditions that are no longer documented get flagged for removal. Conditions that progressed or resolved get flagged for updated coding. The output is a submission set that reflects current clinical reality rather than accumulated historical coding.
Stopping the Drift
Retrospective Risk Adjustment programs that don’t re-validate prior-year codes annually are building a growing gap between their submissions and their members’ actual conditions. That gap is the source of the coding intensity signal CMS monitors, the audit findings OIG publishes, and the settlement amounts DOJ collects. Annual re-validation breaks the carrier-forward cycle, stabilizes risk profiles at clinically accurate levels, and produces submission sets that reflect what’s true now rather than what was true three years ago.
Also,, Read: https://magazinesbeta.com/
