The Structural Redesign of Medicare Risk Adjustment
Part I: The Architecture of V28 – Methodology, Calibration, and Intent
To comprehend the magnitude of the 2026 financial cliff, one must first deconstruct the statistical and clinical foundations of the V28 model. Unlike its predecessor, Version 24 (V24), which was built on legacy frameworks, V28 serves as a "clinical revision" designed to modernize how risk is quantified.
1.1 The Shift from ICD-9 Legacy to ICD-10 Specificity
The V24 model, while operational during the ICD-10 era, was fundamentally anachronistic. It utilized grouping logic based on ICD-9-CM codes, which were then "cross walked" or mapped to ICD-10 codes. This translation layer effectively flattened the clinical granularity available in ICD-10, treating specific diagnoses with the same weight as their unspecified counterparts.
V28 eliminates this translation layer entirely. It is the first CMS-HCC model calibrated directly on the ICD-10-CM classification system.1 This structural shift allowed CMS to re-evaluate the predictive power of every diagnosis code. By analyzing data from 2018 (diagnoses) and 2019 (expenditures), CMS identified thousands of codes that, while frequently used, did not statistically predict increased marginal costs for the Medicare Trust Fund.
Key Structural Changes:
- Expansion of Categories: The model expanded the number of Hierarchical Condition Categories (HCCs) from 86 in V24 to 115 in V28.3 This increase suggests a more nuanced classification system capable of capturing different severities of illness.
- Reduction in Mapped Codes: Despite the increase in HCC categories, the volume of diagnosis codes triggering payment plummeted. V28 removes approximately 2,294 ICD-10-CM codes that previously mapped to payment HCCs in V24.5 Conversely, only roughly 268 new codes were added.
This divergence—more categories but fewer codes—signals a clear policy intent: to reward specificity and penalize ambiguity. Codes that lack clinical detail (e.g., "unspecified") or represent conditions with low cost-predictability (e.g., mild depression) have been systematically purged from the payment model.
1.2 The "Principle 10" and Clinical Reclassification
The restructuring of V28 was guided by "Principle 10," a standard used by CMS to evaluate risk adjustment models. This principle states that discretionary diagnostic categories—those subject to wide variations in coding practice or "gaming"—should be excluded or constrained to prevent upcoding.3
Under this principle, CMS targeted conditions where the clinical criteria were vague or where the diagnosis was often carried over from year to year without active management ("history of" codes). This led to the elimination of high-frequency codes for conditions like Protein-Calorie Malnutrition and Angina Pectoris, which were deemed to be "magnets" for coding intensity rather than indicators of true resource utilization.9
1.3 The Principle of Constraining
Perhaps the most mathematically significant change in V28 is the introduction of "constraining." In V24, hierarchies often provided higher payments for complications. For example, "Diabetes with Acute Complications" carried a higher coefficient than "Diabetes without Complications." This created a perverse incentive to find any complication to elevate the risk score.
In V28, CMS applied constraints to specific disease families, forcing the coefficients (payment weights) to be equal across different severity levels.
- Diabetes Constraint: The three new diabetes HCCs (36, 37, and 38) all carry the same coefficient value.11 Whether a patient has uncomplicated diabetes or diabetes with chronic complications, the base risk score contribution is identical.
- Heart Failure Constraint: A similar logic was applied to Congestive Heart Failure (CHF). While the model distinguishes between acute and chronic failure clinically, the coefficients for HCCs 224, 225, and 226 are constrained to be equal.6
This statistical flattening removes billions of dollars from the system that were previously generated by distinguishing between severity levels, forcing plans to rely on disease interaction terms (discussed in Part VI) to capture true high-risk complexity.
Part II: The Transition Mechanics – The Road to 2026
The implementation of V28 was not instantaneous. Understanding the transition timeline is critical for interpreting financial performance in 2025 and projecting revenue for 2026.
2.1 The Blended Payment Model
To mitigate the immediate financial shock to Medicare Advantage Organizations (MAOs), CMS implemented a three-year phase-in strategy. Risk scores for the transition years are calculated as a blend of the old (V24) and new (V28) models.
Table 1: The CMS-HCC Phase-In Timeline
Source: 1
Strategic Implication:
As of December 15, 2025, the industry is concluding the data collection period for Payment Year 2026. This creates a critical urgency: Every diagnosis captured today will be reimbursed at 100% V28 rates. The safety net of the V24 model, which protected 33% of revenue in 2025, will completely vanish on January 1, 2026. Organizations that have not fully transitioned their documentation practices to V28 standards by now are effectively locking in revenue losses for the upcoming year.
2.2 Normalization Factors and Coding Intensity
Beyond the blend, the raw risk score is subject to adjustments that further compress revenue.
- Normalization Factor: This factor adjusts the raw risk score to account for the trend of rising risk scores over time (population aging and coding improvement). For V28, the normalization factors have generally been lower than V24 (e.g., 1.045 vs 1.153 in 2024), reflecting that the V28 model naturally produces lower raw scores.1
- MA Coding Pattern Adjustment: CMS applies a statutory minimum adjustment (typically 5.9%) to reduce MA risk scores, aligning them with FFS coding intensity. This adjustment is applied to the final blended score, acting as a universal deflator on revenue.1
Part III: Clinical Impact Analysis – The "Losers" (Removed Conditions)
The financial impact of V28 is primarily driven by the removal of high-volume diagnosis codes that previously generated significant revenue. These "losers" in the V28 model represent conditions that CMS deemed to have low predictive value for cost or were prone to discretionary coding.
3.1 Vascular Disease: The Dismantling of HCC 108
In V24, "Vascular Disease" (HCC 108) was a foundational category. It captured a wide range of common atherosclerotic conditions, including "Atherosclerosis of native arteries of the extremities" (e.g., I70.209) and unspecified Peripheral Vascular Disease (PVD) (I73.9). These codes were prevalent in the elderly population and relatively easy to capture.
The V28 Shift:
- Total Removal of Uncomplicated Vascular Disease: In V28, codes for atherosclerosis without severe complications (ulceration or gangrene) do not map to a payment HCC.6
- Impact: A patient documented simply with "PVD" or "Claudication" now generates zero risk adjustment revenue. This effectively wipes out a massive volume of "checkbox" diagnoses.
- New Requirement: To capture risk for vascular disease, documentation must now specify severe complications. The new vascular HCCs (263, 264, 267) are reserved strictly for atherosclerosis with ulceration or gangrene.15 This shifts the focus from chronic management of circulation issues to the management of acute tissue breakdown.
3.2 Metabolic Disorders: The Protein-Calorie Malnutrition Reset
"Protein-Calorie Malnutrition" (HCC 21 in V24) was historically a high-value target for risk adjustment, often documented based on a single low albumin level or vague clinical impressions of "frailty."
The V28 Shift:
- Category Deletion: The broad HCC for Protein-Calorie Malnutrition has been deleted.6 Codes such as E43 (Unspecified severe protein-calorie malnutrition), E44.0 (Moderate), and E46 (Unspecified) generally do not map to a payment category in V28.2
- Clinical Rationale: CMS and the OIG found that these diagnoses were often unsupported by clinical evidence of severe malnutrition (e.g., significant muscle wasting, starvation) in the community setting.
- Exceptions: V28 retains payment for specific, verifiable metabolic disorders such as Morbid Obesity (HCC 48) and rare genetic metabolic conditions (HCC 49, 50, 51).18 However, the ubiquitous "malnutrition" diagnosis is no longer a viable revenue driver.
3.3 Behavioral Health: The Severity Threshold
Mental health coding has undergone a rigorous contraction. Under V24, Major Depressive Disorder (MDD) mapped to a payment HCC regardless of severity—"mild," "unspecified," or "in remission" codes all triggered payment.
The V28 Shift:
- Removal of Mild/Remission Codes: V28 explicitly removes ICD-10 codes for MDD specified as mild (F32.0, F33.0) or in remission (F32.4, F32.5, F33.4).2
- Documentation Mandate: Payment is now reserved exclusively for moderate or severe MDD (e.g., F32.1, F32.2).3
- Strategic Implication: This change demands that providers utilize clinical tools like the PHQ-9 to objectively assess and document the current severity of the disorder. A note stating, "History of Depression, stable on meds" will likely lead to a non-payment code (remission or unspecified), whereas "Major Depression, Recurrent, Moderate; PHQ-9 score 12" supports a payment HCC.
3.4 Angina Pectoris
Angina Pectoris (HCC 88 in V24) was another high-frequency code used for patients with stable coronary artery disease (CAD) who experienced chest pain.
The V28 Shift:
- Category Removal: The HCC for Angina Pectoris has been removed.6
- Clinical Alignment: CMS determined that stable angina is largely a symptom of CAD, which itself is often not a high-weight HCC unless accompanied by acute events (MI). This removes the ability to "stack" angina on top of CAD for additional revenue. The focus shifts to Unstable Angina (if admitted/acute) or the management of the underlying ischemic heart disease, though the revenue potential is significantly lower compared to the V24 model.
Part IV: Clinical Impact Analysis – The "Winners" (New & Expanded Conditions)
While V28 reduced the total number of valid codes, it expanded the number of HCC categories from 86 to 115. This expansion created "winners"—specific, complex conditions that now carry distinct payment weight, rewarding providers for high-fidelity documentation.
4.1 Severe Persistent Asthma (HCC 279)
One of the most significant new opportunities in V28 is the creation of HCC 279: Severe Persistent Asthma.6
- V24 Context: In the previous model, asthma was often grouped into broader pulmonary categories or required a diagnosis of Chronic Obstructive Pulmonary Disease (COPD) to trigger significant value. Asthma alone was often a non-payment condition.
- V28 Opportunity: Recognizing the high cost of managing severe asthma (e.g., biologics, frequent hospitalizations), CMS introduced a dedicated HCC. Specific ICD-10 codes such as J45.50 (Severe persistent asthma, uncomplicated), J45.51 (with acute exacerbation), and J45.52 (with status asthmaticus) now map to HCC 279.19
- Documentation Strategy: The diagnosis must explicitly state "Severe Persistent." "Asthma NOS" or "Chronic Asthma" will typically default to non-payment codes. Providers must document the frequency of symptoms (e.g., continuous throughout the day), night-time awakenings, and interference with normal activity to clinically justify this severity level.
4.2 eating disorders (HCC 153)
V28 introduces specific recognition for Anorexia Nervosa and Bulimia Nervosa under HCC 153.6
- Demographic Shift: While traditionally associated with younger populations, the inclusion of these codes recognizes the clinical reality of disordered eating in the elderly—often secondary to depression, cognitive decline, or medication side effects—as a distinct cost driver separate from general malnutrition.
4.3 Heart Failure Specificity (HCC 221-226)
The Heart Disease group underwent a massive expansion, growing from 5 HCCs in V24 to 10 HCCs in V28.21 This allows for a granular stratification of heart failure that was previously impossible.
- The New Hierarchy:
- HCC 221: Heart Transplant Status/Complications
- HCC 222: End-Stage Heart Failure
- HCC 223: Heart Failure with Heart Assist Device/Artificial Heart
- HCC 224: Acute on Chronic Heart Failure
- HCC 225: Acute Heart Failure (Excludes Acute on Chronic)
- HCC 226: Heart Failure, Except End-Stage and Acute
- Operational Requirement: While some of these categories are constrained (see Part V), the existence of these distinct buckets mandates specificity. Providers must document the type (Systolic, Diastolic, Combined) and the chronicity (Acute, Chronic, Acute on Chronic).23 "CHF Unspecified" (I50.9) is functionally obsolete in a model that demands this level of detail.
4.4 Atherosclerosis with Complications (HCC 263, 264)
While simple PVD was removed, V28 created high-value categories for Atherosclerosis with Ulceration or Gangrene.3
- The Specificity Win: This rewards the capture of the causal link between vascular disease and tissue breakdown. A code like I70.244 (Atherosclerosis of native arteries of left leg with ulceration of heel and midfoot) captures the full clinical picture.
- Documentation nuance: The documentation must link the ulcer to the atherosclerosis. Simply coding "PVD" and "Leg Ulcer" separately may not trigger the high-value combination code unless the provider explicitly links them (e.g., "Arterial ulcer due to PVD").
4.5 Rare Diseases and Metabolic Disorders
V28 added several HCCs for rare but high-cost conditions, reflecting a move toward capturing "outlier" risks.
- HCC 49: Specified Lysosomal Storage Disorders (e.g., Gaucher disease, Pompe disease). This carries a very high coefficient (approx. 6-9x a standard chronic condition).18
- HCC 111: Hemophilia (Male).
- HCC 50: Amyloidosis and Porphyria.
- These additions require organizations to ensure their suspecting logic can identify these rare conditions, which may have been historically under-coded or buried in "other metabolic disorder" categories.
Part V: The Financial Mathematics – Constraining and Interactions
The financial impact of V28 extends beyond the simple addition or subtraction of codes. The re-weighting of the model through "constraining" and "interactions" fundamentally alters the math of risk scoring.
5.1 The Constraining Effect on RAF Scores
Constraining is the process of assigning the same coefficient to multiple HCCs within a disease hierarchy to neutralize the financial incentive for upcoding severity.
Table 2: Diabetes Constraining Impact (Illustrative)
Condition
V24 Handling
V28 Handling
Impact
Diabetes w/ Acute Complications
Higher Coefficient
HCC 36 (Constrained)
Revenue Neutral
Diabetes w/ Chronic Complications
Medium Coefficient
HCC 37 (Constrained)
Revenue Neutral
Diabetes Uncomplicated
Lower Coefficient
HCC 38 (Constrained)
Revenue Neutral
Analysis: In V28, all three diabetes HCCs (36, 37, 38) carry the same coefficient (approximately 0.166 for a community, non-dual, aged beneficiary).11 This creates a "flat" payment curve for diabetes. Documenting "Diabetes with Hyperglycemia" no longer generates more revenue than "Diabetes Type 2." This removes billions of dollars from the system that were previously driven by the aggressive coding of acute complications.
Heart Failure Constraining:
Similarly, HCCs 224 (Acute on Chronic), 225 (Acute), and 226 (Chronic) share the same coefficient (approx. 0.360).6 Despite the clinical difference between a patient in acute decompensation and one with stable chronic heart failure, the base risk score contribution is identical. This places the burden of revenue capture on Interaction Terms rather than severity escalation.
5.2 Interaction Terms: The Multiplier Effect
Because base coefficients for conditions like Diabetes and CHF have been constrained (lowered/flattened), the "Interaction Terms"—extra points added when two conditions coexist—become disproportionately important.
Key Interaction: Diabetes + Heart Failure
- V24: Diabetes + CHF interaction added ~0.121 to the score.
- V28: Diabetes + CHF interaction adds ~0.112 to the score.2
- Strategic Nuance: Even though the base scores for Diabetes and CHF are constrained, the combination remains a massive value driver. Failing to capture one of these conditions (e.g., missing the CHF diagnosis because it wasn't specific enough) causes a "double loss": the loss of the CHF base score plus the loss of the interaction add-on.
Other Key Interactions:
- CHF + Atrial Fibrillation: This interaction remains in V28 (approx. 0.077 add-on).26 Since Afib is a common comorbidity with Heart Failure, ensuring both are captured is vital for accurately reflecting patient complexity.
5.3 Comparative Patient Analysis (V24 vs. V28)
To visualize the financial impact, consider a hypothetical patient profile analyzed in the transition studies.2
Patient Profile: 73-year-old female, Community, Non-Dual.
- Diagnoses: Type 2 Diabetes w/ PVD, Chronic Heart Failure, Protein-Calorie Malnutrition, Morbid Obesity.
Diagnosis
V24 Coding & Impact
V28 Coding & Impact
Result
Diabetes w/ PVD
HCC 18 + HCC 108 (High Value)
HCC 37 (Constrained Value)
Decrease
Heart Failure
HCC 85 (Standard Value)
HCC 226 (Constrained Value)
Neutral/Decrease
Malnutrition
HCC 21 (High Value)
No Map (Removed)
100% Loss
Morbid Obesity
HCC 22 (Standard Value)
HCC 48 (Standard Value)
Neutral
Interaction
DM + CHF + CVD Interactions
DM + CHF Interaction
Decrease
Total RAF Score
~1.736
~1.045
~40% Drop
Source: Derived from 2 examples.
This analysis demonstrates how a stable patient with multiple chronic conditions can see a drastic reduction in risk score purely due to model mechanics (removal of Malnutrition/Vascular and constraining of Diabetes), without any change in actual health status.
Part VI: Operational Strategy – Surviving the 2026 Cliff
With the 100% V28 implementation date of January 1, 2026, imminent, healthcare organizations must pivot from "transition" planning to "execution" mode. The operational focus must shift from volume-based coding to precision-based documentation.
6.1 Clinical Documentation Integrity (CDI): The "MEAT" Mandate
The removal of "history of" and "unspecified" codes places a premium on the MEAT standard (Monitor, Evaluate, Assess, Treat) for every encounter.
- Active vs. Historic: A diagnosis of "Depression" found in the past medical history is worthless in V28. To capture the HCC, the provider must document active management: "Major Depressive Disorder, Moderate. Patient reports mood is stable on current SSRI dose; continue therapy."
- Linking Conditions: For the new "Atherosclerosis with Ulceration" codes, CDI teams must query providers to explicitly link the conditions. "PVD" and "Ulcer" on separate lines of the assessment are insufficient. The note must read "Diabetic ulcer of the foot" or "Ulcer due to atherosclerosis."
6.2 Targeted Specificity Campaigns
Organizations should launch targeted education campaigns for the four high-impact disease groups:
- Vascular: Educate on the removal of uncomplicated PVD. Instruct providers to document gangrene, rest pain, or ulceration if present.
- Depression: Deploy PHQ-9 tools universally to justify "Moderate" or "Severe" coding. Discourage the use of "Depression NOS."
- Asthma: Identify patients on biologics or high-dose steroids. Query providers to update "Asthma" to "Severe Persistent Asthma" where clinical criteria are met.
- Heart Failure: Mandate the documentation of Type (Systolic/Diastolic) and Acuity (Acute/Chronic) for every CHF patient.
6.3 Technology and AI Integration
Manual chart review is insufficient for the complexity of V28.
- Natural Language Processing (NLP): AI tools must be retuned. Algorithms that flagged "Malnutrition" based on low albumin should be deprecated. New algorithms must scan unstructured text (nursing notes, radiology reports) for keywords like "gangrene," "severe persistent," "cachexia," or "ulcer depth" to identify support for V28 codes.
- Prospective Prompting: Electronic Health Record (EHR) alerts must be recalibrated. Prompts for "Angina" or "PVD" are now creating "noise"—distracting providers with non-payment queries. Alerts should focus on missing specificity: "You documented Asthma. Is it Severe Persistent?" or "You documented Depression. Is it Moderate or Severe?"
6.4 Audit Defense (RADV)
The specificity of V28 creates new audit risks. If an organization aggressively codes "Severe Persistent Asthma" (HCC 279) to capture the high revenue, but the medical record only shows a prescription for an occasional albuterol inhaler, that claim is highly vulnerable in a Risk Adjustment Data Validation (RADV) audit.
- Strategy: Ensure that every high-value V28 code is supported by commensurate clinical evidence (e.g., PFT results, biologic prescriptions, PHQ-9 scores) in the chart. Specificity without evidence is a compliance liability.
Part VII: Broader Implications for Value-Based Care
The transition to V28 is not an isolated billing update; it is a signal of the future direction of Value-Based Care (VBC).
7.1 Alignment with True Acuity
By removing "easy" codes, CMS is forcing MA plans to demonstrate value by managing truly complex patients. Revenue is now tied to managing severe organ failure, complications of disease, and severe mental illness, rather than accumulating lists of stable, mild conditions. This aligns payment with the patients who actually incur the highest costs.
7.2 The Bid Strategy Squeeze
Actuaries and finance teams must adjust their bid strategies for 2026. If a plan projects revenue based on V24-era risk scores, they will likely face a revenue shortfall. Plans must anticipate lower aggregate RAF scores and adjust their premiums, benefits, and Medical Loss Ratio (MLR) projections accordingly. The "V28 Cliff" is a solvency risk for plans that have not adequately modeled the 100% implementation impact.
7.3 The Rise of "Specialist" Coding
Primary care providers (PCPs) have traditionally driven risk capture. However, V28's specificity (e.g., detailed Heart Failure types, specific Oncologic diagnoses, rare metabolic disorders) places a higher burden on specialists. MA plans may need to engage Cardiologists, Pulmonologists, and Oncologists more deeply in the risk adjustment documentation process, as they possess the clinical data (Echo reports, PFTs, Pathology) needed to substantiate the new high-value codes.
Part VIII: Conclusion and Strategic Roadmap
As the healthcare industry stands on the verge of 2026, the V28 transition has evolved from a theoretical policy proposal to an operational imperative. The data collection for the first year of 100% V28 payment is happening right now, in the final weeks of 2025.
Summary of Key Actions for December 2025:
- Purge the Problem List: Systematically review and archive diagnosis codes that are valid in V24 but invalid in V28 (e.g., unspecified PVD, mild depression) to prevent them from carrying over into 2026 documentation.
- Re-Train on Specificity: Focus provider education on the "Must-Haves": Heart Failure Type/Acuity, Depression Severity, and Asthma Severity.
- Calibrate Financial Models: Ensure 2026 revenue forecasts account for the removal of the 33% V24 blend.
- Leverage Technology: Use NLP to find the "missing link" between chronic diseases and their acute complications (e.g., Diabetes -> Ulcer) in unstructured notes.
The V28 model represents a tougher, stricter, but ultimately more accurate standard for risk adjustment. Organizations that cling to the volume-based habits of the V24 era will face inevitable financial decline. Conversely, those that embrace the precision of V28—investing in clinical documentation excellence and advanced technology—will find new opportunities to accurately reflect the high acuity of their patient populations and secure the resources necessary to care for them.
The path forward is uncompromising: In the V28 era, specificity is the only currency that matters.
