HCC Hierarchy and Trumping Rules Explained: Which Codes Override Which in V28

The Silent RAF Killer You Might Be Ignoring

You submit a perfectly coded encounter. The ICD-10 codes are valid. The documentation supports them. The HCCs map correctly. But when the RAF score comes back, it is lower than you expected. What happened?

Hierarchies happened.

CMS-HCC hierarchies are a set of rules that automatically zero out lower-severity HCCs when a higher-severity HCC in the same clinical family is present. It is called "trumping" because the higher code trumps the lower one. The lower HCC still exists in the data -- it just contributes nothing to the RAF score.

If you do not understand these rules, you will either waste effort chasing codes that get trumped, or worse, miss the higher-value code that would have survived the hierarchy. Both mistakes cost real money.

This guide covers every major hierarchy group in V28, compares them to V24, and shows you how to build a coding strategy that works with the hierarchy instead of against it.

Why CMS Uses Hierarchies in the First Place

The CMS-HCC risk adjustment model is designed to predict healthcare costs. If a patient has severe heart failure, they are expected to be expensive. If they also have mild heart failure, that does not make them *additionally* expensive -- the severe condition already accounts for the cost.

Without hierarchies, a patient coded with both severe and moderate forms of the same disease would receive credit for both, inflating the RAF score beyond what the data supports. Hierarchies prevent this double-counting by applying a simple rule: within a defined clinical group, only the highest-severity HCC counts toward the RAF score. The HIER files that define every group are published alongside the CMS 2026 risk-adjustment model software and ICD-10 mappings.

This is not optional. It is not something your plan administrator toggles on or off. It is hard-coded into the CMS-HCC model and applied automatically during risk score calculation.

How Trumping Works: The Mechanics

Every hierarchy follows the same pattern:

1. CMS defines a hierarchy group -- a set of HCCs that represent the same clinical condition at different severity levels.

2. The HCCs within the group are ranked from highest severity (highest RAF coefficient) to lowest.

3. When a beneficiary has multiple HCCs within the same hierarchy group, only the top-ranked HCC is counted. All others are set to zero.

Here is a critical nuance: trumping only happens within the same hierarchy group. An HCC in the diabetes hierarchy will never trump an HCC in the heart failure hierarchy. They are independent systems. This means that a patient with both severe diabetes complications and moderate heart failure will receive full credit for both -- because they sit in different hierarchy groups.

The Major V28 Hierarchy Groups

1. Diabetes Hierarchy (HCCs 35, 36, 37, 38)

This is one of the most consequential hierarchies in V28 because diabetes is so prevalent in risk-adjusted populations.

Practical example: A patient has documented diabetes with diabetic chronic kidney disease (mapping to HCC 36) and diabetes without complication (mapping to HCC 38). Only HCC 36 counts. HCC 38 is zeroed out. The RAF coefficient for HCC 38 contributes nothing.

Coding takeaway: Never settle for "diabetes without complication" (E11.9) if the medical record supports a chronic complication. That E11.9 maps to HCC 38, which gets trumped the moment any higher diabetes HCC is present. Document the specific manifestation -- diabetic neuropathy, diabetic retinopathy, diabetic CKD -- and capture the code that survives the hierarchy.

For quick lookups on which ICD-10 codes map to which diabetes HCC, use the HCC Buddy Encoder.

2. Heart Failure Hierarchy (HCCs 224, 225, 226)

V28 restructured heart failure into a cleaner three-tier hierarchy that aligns with clinical severity.

Practical example: A patient has both acute systolic (congestive) heart failure (HCC 224) and chronic diastolic heart failure documented at a lower severity (HCC 226). Only HCC 224 counts.

Coding takeaway: Heart failure specificity matters enormously. Documentation that says "CHF" without specifying type (systolic vs. diastolic), acuity (acute, chronic, acute-on-chronic), or severity will default to a lower HCC that is likely to be trumped. Push for documentation of the most severe, clinically accurate form.

3. Chronic Kidney Disease / Renal Hierarchy (HCCs 326, 327, 328, 329)

CKD is another high-prevalence condition where the hierarchy has significant RAF impact.

Practical example: A patient has Stage 4 CKD (HCC 327) and Stage 2 CKD is also listed on their problem list from a prior year (HCC 329). Only HCC 327 contributes to the RAF score. The Stage 2 code is trumped.

Coding takeaway: Ensure the most current CKD stage is documented. A stale problem list that still shows Stage 2 CKD when the patient has progressed to Stage 4 costs you nothing in the hierarchy (the higher code wins), but the reverse -- failing to document the progression -- leaves significant RAF value uncaptured.

4. Vascular Disease Hierarchy (HCCs 263, 264, 267, 268)

V28 restructured vascular disease significantly compared to V24. Instead of one broad category, there are now distinct hierarchies for different vascular conditions.

Practical example: A patient has both aortic aneurysm with rupture (HCC 263) and documented aortic atherosclerosis without complication (HCC 264). Only HCC 263 counts.

Coding takeaway: Note that HCC 263/264 and HCC 267/268 are *separate* hierarchy pairs. A patient can have both an aortic condition (HCC 263) and a peripheral vascular condition (HCC 268) simultaneously, and both will count because they are in different hierarchy chains. The old V24 approach of lumping all vascular disease together is gone.

5. Cancer Hierarchy (HCCs 21, 22, 23)

Cancer hierarchies reflect the difference between active malignancy and historical cancer.

Practical example: A patient actively receiving chemotherapy for lymphoma (HCC 21) also has a history of a prior, lower-severity malignancy now in remission (HCC 23). Only HCC 21 counts.

Coding takeaway: Active vs. historical status is the key driver. A cancer in active treatment maps much differently than "history of" cancer. Documentation must clearly state whether the malignancy is current, under treatment, or in remission.

6. Substance Use Disorder Hierarchy (HCCs 135, 136, 137)

V28 expanded the substance use hierarchy, which was previously less granular in V24.

Practical example: A patient has opioid dependence with psychotic features (HCC 135) and a separate documented cannabis use disorder (HCC 137). If both map into the same hierarchy group, only HCC 135 counts.

Coding takeaway: Severity qualifiers matter. Document whether the substance use disorder includes dependence, withdrawal complications, or associated psychosis, as these map to higher-severity HCCs.

7. Psychiatric Disorder Hierarchy (HCCs 151, 152, 153, 154, 155)

The psychiatric hierarchy in V28 is more stratified than V24, reflecting the clinical spectrum of mental health conditions.

Practical example: A patient carries diagnoses of both bipolar I disorder (HCC 152) and moderate major depressive disorder (HCC 154). Only HCC 152 contributes to the RAF score.

Coding takeaway: When a patient has co-occurring psychiatric diagnoses that fall within the same hierarchy, the lower-severity condition is trumped. Focus documentation on the highest-severity diagnosis that is clinically accurate and actively managed.

V24 vs. V28 Hierarchy Changes: What Shifted

The transition from V24 to V28 did not just rename HCCs. It fundamentally restructured many hierarchies. Here are the most important shifts.

More hierarchy groups, more granularity. V24 had fewer hierarchy groups with broader categories. V28 broke many of these into finer-grained hierarchies, meaning more opportunities for trumping within each group. The diabetes hierarchy, for example, went from three tiers in V24 (HCCs 17, 18, 19) to four tiers in V28 (HCCs 35, 36, 37, 38).

Vascular disease was completely restructured. In V24, a large number of vascular conditions mapped to a single HCC (108), which has been eliminated. V28 splits vascular disease into aortic (263/264) and peripheral (267/268) hierarchies with severity tiers. This is one of the biggest operational changes for coding teams.

Psychiatric and substance use hierarchies are more complex. V24 had simpler groupings. V28 introduces severity tiers within substance use and psychiatric categories, requiring more specific documentation to capture the highest-value code.

Deleted HCCs cause silent failures. Some V24 HCCs that were commonly coded (like the old vascular disease HCC 108 or protein-calorie malnutrition HCC 47) simply do not exist in V28. If your coding practices have not been updated, you may be submitting codes that generate no HCC credit at all -- not because of trumping, but because the HCC itself is gone.

For a full comparison of V24 vs. V28 changes, see our guide: The Ultimate Guide to CMS-HCC Model V28.

How Hierarchies Interact with Disease Interaction Terms

Hierarchies do not operate in isolation. The CMS-HCC model also includes disease interaction terms -- bonus RAF coefficients that are applied when a patient has specific combinations of HCCs from *different* hierarchy groups.

For example, the V28 model includes interaction terms for combinations like:

Here is the important connection to hierarchies: the HCC that survives the hierarchy is the one that determines whether an interaction term fires. If a patient has diabetes with chronic complication (HCC 36) and diabetes without complication (HCC 38), only HCC 36 survives the diabetes hierarchy. If HCC 36 qualifies for a disease interaction with heart failure but HCC 38 does not, the hierarchy outcome directly affects whether the interaction bonus is applied.

This means that capturing the highest-severity code within each hierarchy group is not just about the individual HCC coefficient. It can also unlock interaction terms that further increase the RAF score.

Impact on RAF Score Calculations: A Worked Example

Let us walk through a concrete RAF calculation to see hierarchies in action.

Patient profile: 72-year-old male, community, non-dual, non-OREC disabled.

Documented conditions:

Before hierarchy application:

All four HCCs are identified: HCC 36, HCC 38, HCC 224, HCC 226, HCC 328.

After hierarchy application:

Surviving HCCs: HCC 36, HCC 224, HCC 328.

Interaction terms: The model checks for qualifying interactions. If Diabetes (HCC 36) + Heart Failure (HCC 224) qualifies for an interaction term, that bonus coefficient is added. If CKD (HCC 328) + Heart Failure (HCC 224) also qualifies, another bonus is added.

The final RAF score is: demographic baseline + HCC 36 coefficient + HCC 224 coefficient + HCC 328 coefficient + applicable interaction bonuses.

Without understanding hierarchies, you might have expected credit for all five HCCs. With hierarchies, you get credit for three -- but those three are the ones that matter, and they unlock interaction terms that the lower-severity codes might not have.

Use the HCC Buddy RAF Calculator to model these scenarios instantly and see exactly which codes survive the hierarchy for any patient.

Common Mistakes and How to Avoid Them

Mistake 1: Coding the lower-severity condition and missing the higher one

This is the most expensive hierarchy mistake. A provider documents "diabetes mellitus without complication" (E11.9, mapping to HCC 38) but the chart clearly shows diabetic neuropathy (mapping to HCC 36). The coder captures E11.9 because it is on the problem list, but never queries for the complication. HCC 38 gets trumped by nothing -- because the higher HCC was never coded in the first place, HCC 38 does contribute. But if a retrospective review later adds HCC 36, the plan discovers that HCC 38 was always going to be redundant.

Fix: Always query for the highest-severity manifestation within each clinical group before finalizing the encounter.

Mistake 2: Spending audit resources on codes that will be trumped

If you already have HCC 224 (severe heart failure) captured for a patient, spending time and resources to also validate HCC 226 (mild heart failure) is wasted effort. HCC 226 will be trumped regardless.

Fix: Prioritize chart review and audit efforts on HCCs that are either the highest in their hierarchy or in hierarchy groups where no code has been captured yet.

Mistake 3: Ignoring hierarchy interactions when prioritizing conditions

Some coders treat all HCCs as equally valuable. But an HCC that triggers a disease interaction term is worth more than its standalone coefficient. Failing to code a qualifying diabetes HCC might not just lose the diabetes coefficient -- it might also lose an interaction bonus with a heart failure HCC.

Fix: Understand which HCC combinations trigger interaction terms, and prioritize capturing codes in those groups.

Mistake 4: Using V24 hierarchy assumptions with V28 codes

The hierarchy structures changed between V24 and V28. An HCC that was at the top of its hierarchy in V24 may not even exist in V28, or may sit at a different tier. Coders who learned hierarchies under V24 need to relearn them for V28.

Fix: Update all reference materials, training documents, and coding tools to reflect V28 hierarchies. Use tools like HCC Buddy that are built natively for V28.

Frequently Asked Questions

What is an HCC hierarchy?

An HCC hierarchy is a rule within the CMS-HCC risk adjustment model that prevents double-counting of related conditions. When a patient has multiple HCCs within the same clinical family (such as two diabetes-related HCCs), only the highest-severity HCC contributes to the RAF score. The lower-severity HCCs are "trumped" and their coefficients are set to zero.

Do hierarchies apply across different disease categories?

No. Hierarchies only apply *within* a defined hierarchy group. A diabetes HCC will never trump a heart failure HCC because they are in completely separate hierarchy groups. A patient can have HCCs from many different groups, and all of them will contribute to the RAF score as long as each one is the highest within its own group.

How did V28 change the hierarchy rules compared to V24?

V28 introduced more hierarchy groups with finer severity tiers. Diabetes went from three HCCs (17, 18, 19) to four (35, 36, 37, 38). Vascular disease was split from one broad category into separate aortic and peripheral hierarchies. Psychiatric and substance use disorders now have more severity levels. Additionally, several V24 HCCs were deleted entirely in V28, meaning some codes that previously had hierarchy relationships now have no HCC mapping at all.

Can a trumped HCC still affect the RAF score through interactions?

No. Once an HCC is trumped within its hierarchy, it is set to zero and does not participate in any disease interaction calculations. Only the surviving (highest-severity) HCC within each group can trigger interaction terms. This is why capturing the highest-severity code is doubly important -- it protects both the standalone coefficient and any interaction bonuses.

What tools can help me identify which codes will be trumped?

The HCC Buddy Encoder allows you to look up any ICD-10 code and instantly see which HCC it maps to, its RAF coefficient, and where it sits in the hierarchy. The HCC Buddy RAF Calculator lets you enter multiple conditions for a patient and see the post-hierarchy RAF score, showing you exactly which HCCs survived and which were trumped. Both tools are built natively for the V28 model.