Medicare risk adjustment is approaching a turning point. In recent years, the Centers for Medicare & Medicaid Services (CMS) has moved beyond incremental rulemaking, signaling a broader shift in how health risk is documented, validated, and potentially inferred.
V28 has completed its three-year phase-in and risk adjustment data validation (RADV) audits now reach more than 550 eligible contracts each quarter, while expanded record submissions are underway for payment years 2020, 2021, and 2024. Beginning in payment year 2027, unlinked chart review records and audio-only diagnoses will be excluded from Medicare Advantage risk scores. CMS also confirmed in its recent LEAD Model FAQ that it is developing an AI-inferred risk adjustment model, with a potential path to full implementation by payment year 2031.
Together, these developments point to two CMS priorities: strengthening the diagnosis-based model in place today while testing whether a new foundation could better support the future. Health plans should prepare their risk adjustment strategy accordingly.
Webinar: Coding guidelines and key updates to prepare for in 2027
Strengthening today’s model while closing its gaps
CMS is preserving the diagnosis-code foundation of risk adjustment while narrowing the ways unsupported, nonspecific, or retrospectively added diagnoses can affect payment. V28 is the clearest example. By reducing the number of valid ICD-10-CM codes, expanding the number of hierarchical condition categories (HCCs), and constraining coefficients, the model places greater emphasis on precise, clinically supported documentation.
CMS’s 2027 policies are no exception. Although CMS paused recalibration using more recent Original Medicare data, it indicated that the pause is temporary. The exclusion of unlinked chart review diagnoses was finalized in the payment year 2027 announcement, and recalibration could return in the payment year 2028 Advance Notice.
At the same time, RADV has shifted from an occasional audit event to an ongoing operating reality. A broader audit cadence, AI-assisted review tools, and the continued pursuit of extrapolated damages create greater operational pressure and financial uncertainty. A late-2025 federal district court decision vacated CMS’s extrapolation methodology and elimination of the fee-for-service adjuster on procedural grounds, but it did not reject extrapolation itself. CMS has appealed, with a Fifth Circuit decision expected in late 2026 or early 2027.
With all of these updates in mind, plans and providers should notice that volume-driven, add-only approaches are becoming increasingly vulnerable. Risk adjustment should now be treated as a clinical documentation and care-delivery discipline rather than simply a coding exercise. Prospective identification, point-of-care support, clinical documentation improvement (CDI), and post-visit pre-bill review can strengthen condition validity and address unsupported or nonspecific coding before claims are submitted.
Testing a new foundation for the future
Through the Long-term Enhanced ACO Design (LEAD) Model, CMS is exploring whether health risk can be inferred from broader data patterns rather than primarily captured through submitted diagnosis codes. Important operational details remain undefined, but the implementation timeline shows a deliberate progression.
- Payment year 2028: CMS will share AI-inferred risk scores with LEAD ACOs alongside traditional HCC scores, without affecting payment.
- Payment year 2029: Payment will blend one-third AI-inferred scores with two-thirds traditional scores.
- Payment year 2030: The weighting will reverse, with two-thirds based on AI-inferred scores.
- Payment year 2031: CMS intends to use 100% AI-inferred scores for the Aged & Disabled population, with possible expansion to High Needs beneficiaries.
The model could address several incentives that CMS has been working to manage, including dependence on chart reviews, coding intensity, coefficient distortions, and audit complexity. Drawing on utilization, pharmacy, laboratory, and other data could reduce reliance on submitted diagnoses and provide a broader view of beneficiary risk.
That does not mean Medicare Advantage risk adjustment will change overnight. LEAD remains a CMS model with guardrails for subgroup performance, prediction accuracy, and benchmark stability. Still, the strategic question is changing from whether CMS wants to address the current incentive structure to how far it may go in replacing it.
Three actions health plans can take now
In preparation for the changes ahead, there are several things that health organizations can do:
1. Make documentation more clinically precise
Documentation should clearly reflect each condition’s presence, status, assessment, and plan of care over time. Whether risk is validated through coding rules or inferred from broader data, a defensible clinical record will remain essential.
2. Move risk adjustment workflows upstream
Retrospective chart chase should evolve toward prospective, encounter-based workflows. Front-end condition identification, embedded documentation support, CDI alignment, pre-bill review, and accurate encounter data can improve performance under today’s model while preparing plans for more data-driven risk inference.
3. Use 2031 as a planning horizon
The 2028–2031 LEAD timeline should be treated as a strategic planning horizon, not a distant policy footnote. If shadow testing succeeds, AI-inferred risk adjustment could become a credible consideration for Medicare Advantage. Organizations investing now in documentation integrity, clinical accuracy, coordinated workflows, and data infrastructure will be better positioned than those built primarily around coding volume.
CMS’s direction is increasingly consistent: simpler incentives, stronger validation, greater competitive neutrality, and a more accurate view of beneficiary health risk. V28 refinements, chart review restrictions, and expanded RADV audits are reinforcing the current framework, while the LEAD Model tests what may come next. Health plans do not need every detail to act. But building clinically grounded documentation, accurate encounter data, and stronger analytic capabilities today can support compliance now and readiness for the future.
Register for our upcoming Risk Adjustment Essentials webinar
Early preparation creates an opportunity to improve accuracy, tighten compliance, and build a more resilient risk adjustment program. Join us Thursday, October 22 to align teams, technology, and strategy for success in 2027 and beyond.
