HCC Coding and Risk Adjustment: A Beginner’s Guide

If you work in medical coding or revenue cycle management long enough, you’ll eventually run into a term that doesn’t behave like the CPT and ICD-10-CM codes you’re used to: HCC. Hierarchical Condition Category coding isn’t about billing a single visit — it’s about painting an accurate, complete picture of a patient’s health status for an entire year, because that picture directly determines how much a health plan gets paid to care for them. Get it wrong, and either the plan is underfunded for a genuinely sick patient, or — the version regulators worry about far more — a plan gets paid for conditions that were never properly documented at all.

What HCC coding actually is

HCC stands for Hierarchical Condition Category. It’s the diagnosis-based methodology the Centers for Medicare & Medicaid Services (CMS) uses to predict how much a person’s healthcare is likely to cost in the coming year, so that payments to health plans reflect the actual health status of the people they cover — not a flat, one-size-fits-all rate per enrollee. The idea is straightforward even though the mechanics get complicated: a plan covering a population with more chronic illness should be paid more than a plan covering an equally-sized but healthier population, and HCC coding is the data layer that makes that possible.

Under the hood, specific ICD-10-CM diagnosis codes map to specific HCCs. Each HCC a patient has — combined with demographic factors like age, sex, and Medicaid or disability status — adds a coefficient to that patient’s overall Risk Adjustment Factor (RAF) score. CMS uses the RAF score to adjust the monthly payment made to the Medicare Advantage plan covering that person.

CMS-HCC vs. HHS-HCC: two different models

This is a point that trips up a lot of people new to risk adjustment: there isn’t just one HCC model. CMS-HCC applies to Medicare and Medicare Advantage. A separate model, HHS-HCC, applies to ACA marketplace plans. They’re related in concept but differ in real ways:

• Population: CMS-HCC covers Medicare/Medicare Advantage enrollees (primarily 65+, plus qualifying disabled individuals). HHS-HCC covers all-ages ACA marketplace enrollees, including children.
• Timing: CMS-HCC is prospective — this year’s documented diagnoses affect next year’s payment. HHS-HCC is generally described as concurrent, affecting risk-transfer calculations within the same plan year.
• Scope: HHS-HCC includes pediatric and maternity risk categories and prescription-drug-based risk categories that the Medicare-focused CMS-HCC model doesn’t need, since Medicare Advantage doesn’t cover newborns or maternity care the same way.

If you’re coding for a Medicare Advantage plan and for an ACA marketplace plan in the same role, don’t assume the rules are identical — check which model applies to the population in front of you.

How RAF scores connect to payment (without the guesswork)

RAF scores are additive: a member’s total score is the sum of demographic-factor coefficients (age, sex, Medicaid eligibility, disability status, institutional status) plus a coefficient for each qualifying HCC they carry, plus disease-interaction terms the model defines for certain combinations (for example, diabetes plus congestive heart failure can add extra points beyond what the two conditions would contribute separately).

The American Academy of Family Physicians offers a helpful illustration of the range this can cover: a relatively healthy 68-year-old woman with type 2 diabetes and hypertension might have a RAF around 0.428. An otherwise similar patient whose chart also properly documents diabetic polyneuropathy, congestive heart failure, and morbid obesity could reach a RAF around 1.327. (Treat these as illustrative, not as current published coefficients — the point is the direction and scale, not the exact decimals.)

One thing worth being honest about: there’s no clean, universal “dollars per RAF point” formula you can quote. The actual payment impact of a given RAF score depends on each plan’s county-level payment benchmark, star-rating bonus, and bid — genuinely complex, and it varies by plan and year. The reliable, simple version for a beginner to hold onto is: higher RAF score → higher predicted cost → higher capitation payment to the plan, assuming every contributing diagnosis is properly documented and supported.

The V24 → V28 model change (and why 2026 matters)

CMS periodically updates the CMS-HCC model itself, and the current transition is significant enough to know about. The newer model, V28, includes 115 payment HCCs, versus 86 in the prior V24 model — a real expansion and restructuring of categories, partly driven by moving fully onto ICD-10-CM-based mappings.

CMS phased the new model in gradually rather than switching all at once, per its 2024 Rate Announcement:

• 2024: 67% of a member’s risk score came from the prior (V24) model, 33% from the new V28 model.
• 2025: 33% V24, 67% V28.
• 2026 (this year): 100% V28.

That means 2026 is the first year the model runs fully on V28 with no blending — worth flagging for anyone whose risk-adjustment reporting or software still references the older category structure. CMS had originally projected roughly a 3% reduction in average risk scores from the V28 changes alone, though the actual 2024 impact came in closer to about 1% once the phase-in blending was factored in.

Documentation: the MEAT standard

This is the part that most directly affects how a provider needs to write a note, and it’s the single most important documentation concept in risk-adjustment coding. A chronic-condition diagnosis code, listed by itself with no other clinical context, is not considered supported for risk-adjustment purposes. Official coding guidance treats a simple list of diagnoses — with no accompanying assessment or plan — as insufficient.

Instead, each condition being coded for HCC purposes should be supported by at least one of four elements in the encounter note, remembered by the acronym MEAT:

• Monitor — signs, symptoms, or disease progression/regression
• Evaluate — test results, medication effectiveness, response to treatment
• Assess/Address — ordered tests, clinical discussion, chart review, counseling
• Treat — medications, therapies, or other treatment modalities

In practice, this means a coder reviewing a chart for HCC purposes is checking not just “is diabetes on the problem list,” but “did this encounter’s documentation actually monitor, evaluate, address, or treat the diabetes.” If it didn’t, that diagnosis generally shouldn’t be coded for that encounter.

The annual recapture requirement

Here’s a detail that catches a lot of practices off guard, especially ones used to fee-for-service coding conventions: HCC-eligible risk scores reset every calendar year. Chronic conditions do not automatically carry forward. A patient’s congestive heart failure, COPD, or diabetes has to be re-documented — assessed or addressed per the MEAT standard — at a face-to-face encounter during each calendar year, or that HCC drops off and stops contributing to the following year’s risk score, even though the patient obviously still has the condition.

This is why risk-adjustment coding programs put so much emphasis on annual wellness visits and structured chart reviews: without a deliberate, disciplined process to make sure every active chronic condition gets properly re-addressed and documented each year, a practice’s patient population can look — on paper — steadily healthier than it actually is, and payment can quietly fall out of step with real patient need.

Specificity matters more here than almost anywhere else in coding

Vague or unspecified diagnosis codes may not map to a payment HCC at all, or may map to a lower-value category than the patient’s actual condition warrants. Coding to the highest level of specificity supported by the documentation — type and underlying cause, control or severity status, site and laterality, relevant comorbidities and risk factors — isn’t just good coding practice in general; in risk adjustment, it can be the difference between a diagnosis counting toward the patient’s risk score at all.

Compliance: where the real scrutiny is right now

Risk adjustment is one of the most actively audited areas in Medicare Advantage today, and it’s worth knowing the current landscape rather than treating this as an abstract “be careful” warning.

A federal Office of Inspector General (OIG) report issued in October 2024 estimated that $7.5 billion in 2023 risk-adjusted payments were based on diagnoses that appeared only on health risk assessments (HRAs) or HRA-linked chart reviews, with no other supporting service — no follow-up visit, procedure, test, or supply — anywhere else in that patient’s records for the year. Just 20 Medicare Advantage companies accounted for 80% of that total, and about 1.7 million enrollees had diagnoses reported this way. Nearly two-thirds of the flagged amount came specifically from in-home HRAs and HRA-linked chart reviews, which are often administered by the plan or a vendor rather than the patient’s own treating physician.

More recently, an OIG audit released in mid-2026 estimated CMS potentially overpaid Medicare Advantage organizations $462 million for payment year 2021 based on acute stroke diagnosis codes that weren’t supported by medical records — in that audit’s sample of 97 enrollees, every single one lacked a corresponding acute stroke diagnosis anywhere in their inpatient or outpatient hospital records for the year.

There’s also an important, still-unfolding legal development worth knowing about rather than assuming is settled: CMS finalized a rule in February 2023 allowing Risk Adjustment Data Validation (RADV) audits to use statistical extrapolation — projecting the error rate found in a small sample across a plan’s entire enrolled population to calculate overpayment liability — while eliminating a prior adjustment that had accounted for unsupported diagnoses baked into the original fee-for-service data used to build the risk model. In September 2025, a federal court ruled that rule procedurally invalid and vacated it. As of this writing, this is a developing legal situation, not a resolved one — CMS’s options going forward (appeal, revise, re-propose) hadn’t played out. If you work in risk-adjustment compliance, this is worth tracking rather than treating as settled either way.

The practical takeaway for anyone doing this work: document supported diagnoses carefully, don’t let HRA-only findings substitute for real clinical follow-through, and expect this area to stay under active regulatory attention for the foreseeable future.

Who does this work, and how it fits a coding career

Risk-adjustment coding is typically performed by credentialed coders working for Medicare Advantage health plans, physician practices and medical groups participating in MA or ACO arrangements, third-party risk-adjustment vendors, or health-system coding-compliance departments — often reviewing charts retrospectively specifically to confirm that HCC-eligible diagnoses are both coded and properly MEAT-supported.

AAPC offers a dedicated credential for this specialty, the Certified Risk Adjustment Coder (CRC) — AAPC’s own certification page states that credentialed CRCs earn more than non-credentialed coders, though that’s AAPC’s marketing claim about its own credential rather than an independently verified salary statistic, so treat it accordingly rather than as a guaranteed outcome. AHIMA offers a parallel Risk Adjustment Coding microcredential, with an exam covering risk-adjustment methodology, documentation and compliance (including MEAT and RADV audit concepts), medical terminology, and quality/regulatory topics like Medicare Advantage Star Ratings and MIPS.

If you’re already working toward a coding certification and want a specialty with real, sustained demand, risk-adjustment coding is a reasonable one to have on your radar — it sits squarely at the intersection of coding accuracy, compliance, and the parts of healthcare payment that regulators are actively watching.

Undercoding matters in risk-adjustment work too, not just fee-for-service billing — see our guide to running an internal medical coding audit for a practical methodology that applies to both.