The Encounter-Based Risk Adjustment Future

Three articles in this series have traced a single trajectory. MCR-02.03 documented the V28 model reform that recalibrated what diagnoses are worth. MCR-02.02 examined the chart review exclusion that eliminates one mechanism for capturing those diagnoses. This article addresses the endpoint both reforms are building toward: a risk adjustment system where only provider-attested encounter data counts for payment.

CMS has not published a proposed rule for encounter-based risk adjustment. There is no implementation date. But the trajectory is unmistakable, and the agency has been explicit about the direction. The CY 2026 Advance Notice Fact Sheet stated that CMS had been working to calibrate the risk adjustment model using MA encounter data and would have the option to begin phasing in an encounter-data-based model as early as CY 2027. The CY 2027 Advance Notice did not propose that phase-in, but it did propose the chart review exclusion, which functions as a structural precursor: remove the non-encounter data sources one at a time, and what remains is encounter-based RA by subtraction.

CMS Director of Medicare Chris Klomp framed the agency’s three principles for risk adjustment after the advance notice release: simplicity (a single data source reduces auditing complexity), competition neutrality (encounter-based RA eliminates the advantage that aggressive coding operations confer over clinical documentation approaches), and payment accuracy (encounter data tied to provider visits is a more reliable indicator of clinical complexity than retrospective chart extraction). Those principles describe a system that does not yet exist in rule but is being constructed through incremental policy action.

The question for every MA stakeholder is not whether encounter-based RA will arrive. It is what the system will look like when it does, what remains uncertain in the transition, and what organizational preparation looks like now.

What Encounter-Based RA Means

#

The current risk adjustment data architecture accepts diagnoses from multiple submission pathways. The Risk Adjustment Processing System (RAPS) is the legacy pathway where plans submit diagnosis codes in a streamlined format. The Encounter Data Processing System (EDPS) is the newer pathway where plans submit full encounter records that include the diagnosis, the provider, the date of service, and the service rendered. Chart review records, both linked and unlinked, provide a third pathway for diagnosis submission. The current system accepts diagnoses from all three sources and uses them to calculate risk scores. A diagnosis generates payment regardless of whether it arrived through an encounter record, a RAPS submission, or a chart review extraction.

Encounter-based RA would narrow the data architecture to a single source. Only diagnoses submitted through encounter data tied to a documented provider visit would count for risk score calculation. RAPS-only submissions would no longer generate risk-adjusted payment. Chart reviews that produce diagnoses not linked to a qualifying provider encounter would be excluded, which the CY 2027 chart review exclusion already implements for the unlinked subset. The provider encounter would become the authoritative unit of payment-eligible clinical information: a diagnosis counts if and only if a provider assessed and documented it during a billable, face-to-face visit.

The shift is not purely about data format. It is about the evidentiary standard for what constitutes a payment-eligible diagnosis. Under the current system, a plan can identify a diagnosis through retrospective record review and submit it for payment even if no provider evaluated the patient for that condition in the relevant period. Under encounter-based RA, the diagnosis must emerge from the clinical encounter itself. The provider must have seen the patient, assessed the condition, and documented it in the context of the visit. The encounter is both the clinical event and the payment trigger.

For standard MA plans (non-PACE), RAPS has already been largely supplanted by encounter data submission for most risk adjustment purposes, though RAPS data continues to supplement encounter data in certain contexts. The PACE transition is further behind: CMS proposed a 50/50 blend for CY 2027 between the legacy 2017 CMS-HCC model (which incorporates RAPS, encounter data, and FFS claims) and the proposed 2027 model (which relies on encounter data and FFS claims only), with a tentative full transition target of CY 2029. The PACE timeline provides one indicator of the pace at which CMS moves toward encounter-only data for the broader MA population.

The encounter-based RA trajectory also intersects with Risk Adjustment Data Validation (RADV) audits. CMS has been expanding its RADV audit program, committing in May 2025 to audit all eligible MA contracts each payment year and beginning to apply extrapolation of audit findings starting with Payment Year 2018. Under encounter-based RA, RADV audits would validate a single data source rather than reconciling diagnoses across multiple submission pathways. The audit infrastructure becomes simpler and the evidentiary chain between diagnosis and payment becomes more transparent. Plans that submit diagnoses through encounters have a clearer documentation trail linking the diagnosis to a specific provider, visit date, and clinical context. Plans that currently rely on chart review submissions to supplement encounter data face a more complex audit exposure because the documentation trail runs through a retrospective coding process rather than through the clinical encounter itself.

The operational infrastructure required for encounter data submission is also more demanding than RAPS submission. Encounter data records must include the provider, date of service, place of service, procedure codes, and diagnosis codes tied to that service. RAPS submissions require only the diagnosis and a date range. The additional data elements in encounter records create a richer dataset for payment validation but also require plans to maintain more complete and accurate encounter data submission pipelines. Plans with legacy technology infrastructure or fragmented provider data systems will face capital investment requirements to build the encounter data quality that an encounter-only system demands.

The Provider as Gatekeeper

#

Encounter-based RA fundamentally restructures the relationship between plans and providers in risk adjustment.

Under the current system, plans are the active agents in risk capture. A plan’s coding team, chart review vendors, and CDI programs can supplement what providers document during encounters by retrospectively identifying additional diagnoses from the medical record. The plan controls a significant portion of the risk score generation process through its own operational infrastructure, independent of what the provider submitted during the visit.

Under encounter-based RA, the provider’s clinical encounter becomes the sole source of risk adjustment data. The provider is no longer one of several inputs into the risk score. The provider is the gatekeeper. A plan’s revenue depends entirely on what its contracted providers document during visits. Providers who consistently under-document, whether due to time pressure, unfamiliarity with HCC mapping rules, or clinical workflows that prioritize assessment over coding, will cost their contracted plans revenue. Providers who document comprehensively and accurately at the point of care become more valuable contracting partners.

This power shift has practical implications for provider organizations. Documentation accuracy transitions from a compliance function to a revenue-determinative one. Coding completeness at the point of care becomes the primary variable in plan revenue generation, not something that can be corrected after the fact through retrospective review. Provider organizations that invest in clinical documentation training, EHR-integrated coding decision support, and workflow redesign for documentation completeness will be positioned to retain and attract plan contracts. Those that do not will find themselves generating lower risk scores for the same clinical populations, making them less attractive to plans that can contract with higher-documenting provider groups.

The incentive intensity problem follows directly. Plans have a strong financial incentive to ensure providers capture every clinically appropriate HCC at every visit. The CDI program evolves from retrospective chart review, which encounter-based RA eliminates, to prospective clinical documentation support, which encounter-based RA demands. But the line between provider education and coding pressure is not bright. Pre-visit planning tools that flag conditions the patient has been previously documented for and may need reassessment are one thing. EHR-based prompts that direct providers toward specific coding outcomes during the visit are another. AI-driven clinical decision support systems that surface HCC opportunities in real time during the encounter present both the greatest documentation efficiency potential and the greatest compliance risk.

The contracting implications cascade through the provider network. Plans will increasingly evaluate providers not only on cost efficiency and quality metrics but on documentation performance measured as the gap between clinical complexity and coded risk score. A provider group serving a population with high chronic disease burden but generating risk scores below the expected level for that clinical profile is, from the plan’s perspective, leaving revenue on the table. Under the current system, the plan can recover some of that revenue through chart review. Under encounter-based RA, the revenue simply does not exist unless the provider captures it at the point of care. Plans will gravitate toward provider organizations that demonstrate documentation completeness as a measurable capability, and provider compensation models will increasingly incorporate documentation performance metrics.

This dynamic creates tension within primary care in particular. Primary care physicians managing panels of 1,500 to 2,500 patients with multiple chronic conditions already face visit-time constraints. Adding the expectation that every visit captures every clinically relevant HCC at the specificity level the risk adjustment model requires is a documentation burden that competes with clinical assessment, patient communication, and care coordination. The question is whether EHR-based clinical decision support can bridge that gap without converting the provider visit into a coding exercise. The answer will depend on how the technology is designed and deployed, which returns to the compliance question.

The question CMS has not fully addressed, and that plans, providers, and OIG will contest for years, is where legitimate prospective CDI ends and impermissible coding direction begins. The chart review exclusion (MCR-02.02) established a compliance boundary for retrospective activity. Encounter-based RA will require a comparable boundary for prospective activity, and that boundary does not yet exist in regulation.

The Coding Compliance Frontier

#

Encounter-based RA does not eliminate coding compliance risk. It relocates it from the plan’s retrospective coding operations to the plan-provider interface during the encounter.

Legitimate plan-provider coding interactions under encounter-based RA would include educating providers on HCC mapping rules and documentation specificity requirements, providing data on coding gap rates that show providers which conditions they treat but do not consistently document, offering CDI training focused on medical record accuracy and completeness, and deploying pre-visit planning tools that flag conditions the patient has and that may need clinical assessment at the upcoming visit. These activities improve documentation accuracy without substituting the plan’s judgment for the provider’s clinical assessment.

The line is crossed when the plan’s intervention directs the coding outcome rather than improving the documentation process. Presenting providers with pre-populated diagnosis lists for “confirmation” without independent clinical assessment substitutes the plan’s analytics for the provider’s judgment. Financial incentives that tie provider compensation to HCC capture rates rather than clinical quality create a payment structure where the provider is rewarded for coding activity, not for clinical care. Audit pressure on providers whose coding patterns produce lower risk scores than peers imposes a coding productivity standard that has nothing to do with the patient in front of the provider. Using AI or algorithmic tools to direct provider behavior toward specific coding outcomes during the visit, rather than to support the provider’s own clinical documentation, converts a decision support tool into a coding direction mechanism.

The enforcement trajectory makes the compliance stakes concrete. DOJ’s FCA enforcement priority in MA risk adjustment is well established (MCR-02.02). The relaunched DOJ-HHS FCA Working Group identified MA risk score integrity as its first priority lane. OIG’s ongoing series of targeted audits of documentation supporting specific diagnosis codes covers 11 projects, with several completed in 2024 and 2025. The Anti-Kickback Statute implications of coding-linked financial incentives to providers add a second enforcement vector: if a plan pays a provider more for capturing specific diagnoses, and those diagnoses drive risk-adjusted payment from a federal program, the payment structure may create AKS exposure regardless of whether the documented diagnoses are clinically accurate.

Plans that build encounter-based documentation programs will need compliance architectures that demonstrate the provider’s clinical independence at the point of care. The documentation must emerge from the provider’s assessment, not from the plan’s analytics pipeline. The plan can support the provider’s documentation capacity. It cannot direct the provider’s documentation output (see MCR-04.10 for the compliance infrastructure).

The Payvider Structural Advantage

#

The payvider model occupies a structurally advantaged position in an encounter-based RA world, for the same reasons it occupies a structurally advantaged position under the chart review exclusion.

When the plan and the delivery system are the same entity, the plan-provider coding interface does not exist as an external contracting relationship. Documentation standards are internal operational policy, not negotiated contract terms. EHR systems are owned by the organization and can be configured to support documentation completeness without raising the compliance questions that external plan-directed coding prompts create. The compliance question shifts from managing an arm’s-length relationship to governing internal institutional practices, which is a simpler regulatory problem.

Kaiser Permanente, UPMC, Geisinger, and CareOregon capture diagnoses through encounters with their own clinicians. When encounter-based RA arrives, these organizations do not need to restructure a vendor ecosystem or renegotiate provider contracts. Their risk scores already derive from encounter data because their care model produces encounter data as a natural byproduct of clinical operations. The transition cost is minimal because the operational model is already aligned with the regulatory destination.

For standalone insurers that rely on independent provider networks, encounter-based RA requires building the CDI infrastructure, provider education programs, EHR integration capabilities, and compliance oversight that payviders already possess as inherent features of their organizational design. The investment required is substantial and ongoing. National carriers like UnitedHealth, Humana, and CVS/Aetna will need to redirect capital from the chart review infrastructure they are now winding down toward prospective documentation support systems they have not historically needed at scale. Regional nonprofits with closer provider relationships and smaller networks may find the transition less operationally complex but still capital-intensive. The competitive gap between integrated and non-integrated plans widens as the data architecture narrows, because the integrated model’s cost of risk score generation through encounters is structurally lower than the standalone insurer’s cost of achieving the same outcome through contracted provider relationships.

UnitedHealth’s Optum division presents an interesting hybrid case. Optum Health employs or affiliates with tens of thousands of physicians, giving UnitedHealth partial payvider characteristics for a portion of its MA population. The encounter-based RA transition may accelerate UnitedHealth’s strategic investment in Optum Health as a direct care delivery platform, not because of the clinical integration benefits alone but because provider employment gives the plan direct control over documentation workflows in a way that contracted provider relationships do not. Other national carriers without comparable delivery system assets face a starker choice: invest in provider network management capabilities that approximate the payvider documentation advantage, or accept structurally lower risk scores and the revenue gap that follows.

The strategic implication is clear. Encounter-based RA strengthens the business case for delivery system integration. Plans considering the payvider pathway (MCR-05.02) should factor the encounter-based RA trajectory into their strategic calculus. The chart review exclusion has already shifted the competitive landscape. Encounter-based RA would accelerate that shift.

Timeline and Preparation

#

CMS has not committed to a specific implementation date for encounter-based RA. The agency has stated that it would have the option to begin phasing in an encounter-data-based model as early as CY 2027, but the CY 2027 advance notice took the intermediate step of excluding unlinked chart reviews rather than proposing the full transition. This sequencing suggests a multi-year approach: remove unlinked chart reviews first, allow the industry to adapt, then narrow further to encounter-only data in a subsequent payment year.

The rulemaking vehicle is uncertain. Encounter-based RA could come through the annual advance notice and rate announcement process, as the chart review exclusion did, or through a separate notice of proposed rulemaking that would provide a longer comment period and more detailed impact analysis. The magnitude of the change, affecting every MA plan’s risk score generation infrastructure, arguably warrants dedicated rulemaking. But CMS has demonstrated through the chart review exclusion that it is willing to make consequential payment methodology changes through the advance notice process. Industry stakeholders, including AHIP and individual plans, will argue that the advance notice process does not provide adequate time for the financial modeling and operational planning that a transition of this scale requires.

The political dynamics may influence timing as much as the technical readiness. Encounter-based RA would produce another significant reduction in MA plan payments, following the V28 phase-in, the chart review exclusion, and the ongoing coding pattern adjustment. The cumulative impact of sequential payment reductions on plan benefits, service areas, and enrollment creates political exposure for the administration. Whether CMS pushes encounter-based RA in CY 2028 or defers it to CY 2029 or beyond may depend on the outcome of the CY 2027 rate cycle and its visible effects on beneficiary benefits in the midterm election year.

Implementation capacity is a binding constraint. CMS is simultaneously running or launching WISeR, AHEAD, ACCESS, BALANCE, MAHA ELEVATE, and multiple other CMMI models. The regulatory bandwidth required to finalize encounter-based RA, build the auditing infrastructure, manage the transition, and respond to the inevitable litigation is substantial. Whether CMS has the workforce and contractor capacity to execute encounter-based RA while managing the rest of its reform portfolio is an open question (see MCR-03.05 on CMS implementation capacity).

What plans should be doing now does not depend on knowing the implementation date. The preparation requirements are the same whether encounter-based RA arrives in CY 2028 or CY 2030. Plans should audit their current risk score generation to determine what proportion of their HCC revenue derives from encounter data versus chart reviews and RAPS submissions. That proportion is the measure of exposure. Plans should build or upgrade prospective CDI programs oriented around point-of-care documentation support. Provider contracts should be evaluated for encounter-based documentation incentives that reward coding completeness without creating AKS exposure. Scenario modeling should quantify what happens to risk scores and revenue if only encounter-linked diagnoses count for payment. Plans that find a significant gap between their encounter-based and total risk scores have a measurable dollar amount of revenue at risk in the encounter-based RA transition.

Encounter-based risk adjustment is not a policy proposal. It is a policy trajectory. The V28 model recalibrated what conditions are worth. The chart review exclusion eliminated one mechanism for capturing those conditions. Encounter-based RA, when it arrives, will make the provider encounter the sole authoritative source of payment-eligible diagnoses. Plans that wait for a final rule to begin preparing are already behind. The organizations best positioned for this transition are those that recognized, years ago, that risk adjustment should emerge from clinical care rather than administrative coding operations, and built their systems accordingly.

Related Reading

#

MCR-05_01 The Provider’s New Reality: Revenue, Authorization, and Accountability MCR-05_02 Becoming a Payvider: The Strategic Case for Provider Plan Ownership MCR-06_02 BGM and CGM in the Medicare Ecosystem: The Policy Landscape for Glucose Monitoring Vendors