Impact of Unlinked Codes and the Evolution of Retrospective Reviews

Historically, Medicare Advantage plans have benefitted from a broad approach to retrospective coding that allowed for the addition of new risk adjustment codes as long as documentation standards were met. But according to the Centers for Medicare and Medicaid Services (CMS) 2027 final rule, which proposes to exclude diagnosis information from risk scores unless it is tied to a covered encounter, these existing models will no longer suffice, with the exception of some members who switched Medicare Advantage plans.

The financial consequences are expected to be significant, including the recovery of $4.7 billion in overpayments in 2023. While this move is not a surprise for most in the industry, it is coming earlier than expected, and health plans must get their house in order quickly, especially when it comes to chart review.

With this move, CMS has increased the importance of strong chart review by emphasizing the need for organizations to better link diagnoses with care episodes. This will be especially important for MA plans that have relied heavily on unlinked chart review records in the past.

Bottom line: Organizations without a comprehensive and accurate retrospective review should prepare now for the changes now, or risk falling behind.

Which MA plans face the greatest risk?

Healthcare organizations facing the greatest potential losses from this new proposal share common characteristics:

• Dependence on retrospective chart review disconnected from encounter or claims context
• Limited ability to trace suspected conditions to documented provider action
• Vendor strategies optimized for output volume rather than substantiation

If CMS no longer includes diagnoses from unlinked charts in risk score evaluation in the future, MA plans relying on multiple pass reviews will increasingly see diminishing returns for each sweep associated with risk adjustment.

Preparing for the future of risk adjustment

Forward-thinking healthcare organizations will shift strategy related to retrospective chart review from that of a growth lever to a precision exercise that focuses strongly on code validation instead of simple code capture. Future success now hinges on the ability to systematically exclude charts with no tie to a documented encounter and concentrate effort where evidentiary support is strongest.

In addition, MA plans must lean into a stronger first-pass review to create economies of scale. This way, resources can be better invested in prospective reviews and point‑of‑care accuracy to improve compliance and overall performance.

Focus areas include:

• Utilizing technology to confidently prioritize what charts to review, and which to skip
• Complete code capture for retrospective reviews using technology and human-in-the-loop to advance both efficiency and accuracy
• Concurrent and prospective chart reviews to understand a full picture of care

The evaire platform was purpose-built in anticipation of this shift to deliver more nuanced and validated outcomes. An end-to-end AI-powered chart review, coding and analysis solution, evaire improves risk adjustment and quality initiatives by embedding linkage requirements, evidentiary rigor and confidence-driven prioritization directly into risk adjustment workflows.

Powered by agentic AI, MA health plans can count on ECLAT’s proprietary Confidence Score to provide diagnosis-level documentation validation—proactively identifying unsupported ICD-10 codes and assigning a confidence score that notes the probability a diagnosis lacks adequate documentation support. The result is a quantitative metric indicating which records contain new codes or documentation for review, and which to skip, within 99%+ accuracy.

evaire approaches retrospective review from a viewpoint of clinical documentation review, claim validation and compliance. As the industry shifts away from traditional retrospective economics toward encounter validation, ECLAT is one step ahead and ready to help your organization navigate the coming changes.

For more information, reach out to our team of experts here.