Diabetes x heart failure
+0.112
- HCCs 35 to 38 with HCCs 221 to 226, Community NonDual Aged [3]
Guide · Risk adjustment
RAF score explained: how it is calculated for 2027
Short answer: five steps turn a chart into a Medicare Advantage payment: demographic factor, HCC coefficients, hierarchies and interactions, normalization by 1.079, coding adjustment of 5.90%. Worked member below: 0.395 to 1.503, on the 2024 CMS-HCC model with the CY 2027 factors [2].
2024 CMS-HCC model · V28115 payment HCCs2026 DOS build the 2027 score1.079 · 5.90%
Focus
RAF arithmetic
Framework
2024 CMS-HCC (V28)
Codes
115 payment HCCs
Metric
1.079 · 5.90%
Archetype
72-year-old, 3 HCCs
Length
8 min
From chart to payment: five steps
A Risk Adjustment Factor score, close to 1.0 on average, scales what Medicare pays an MA plan for one enrollee for one year. It is prospective: 2027 uses 2026 diagnoses, so a 2025 diagnosis not documented at a 2026 visit contributes nothing [2].
- 1Demographicssegment factorAge, sex, Medicaid and disability status pick the column
- 2HCCsadd coefficientsEach documented condition maps to one of 115 categories
- 3Hierarchy and interactionssuppress, then addOnly the most severe in a family pays; pairs add weight
- 4Normalizedivide by 1.079Holds average FFS to 1.0 [2]
- 5Coding adjustmentreduce 5.90%Uniform across plans for 2027 [2]
The demographic component
The segment decides every number that follows; dual eligibility, disability and institutional status have their own columns. Table VIII-1, CY 2024 Rate Announcement [3]. A 72-year-old woman with no documented conditions carries a raw 0.395.
| Variable, Community NonDual Aged | Relative factor |
|---|---|
| Female, 65 to 69 years | 0.330 |
| Female, 70 to 74 years | 0.395 worked example |
| Female, 75 to 79 years | 0.465 |
| Male, 65 to 69 years | 0.332 |
| Male, 70 to 74 years | 0.396 |
| Male, 75 to 79 years | 0.502 |
The HCC coefficients
Two things jump out: diabetes with and without complications carries the same weight, so complication coding adds precision but not score; and kidney stage 4 is worth four times stage 3, which is why the note has to state the stage.
Selected payment HCC coefficients, Community NonDual Aged, 2024 CMS-HCC model [3]
Relative factors before normalization; the same table applies for 2026 and 2027 payment because the model did not change [2].
HCC 327Chronic kidney disease, severe (stage 4)
0.5144x stage 3
HCC 226Heart failure, except end-stage and acute
0.36070% of stage 4
HCC 37Diabetes with chronic complications
0.166same as HCC 38
HCC 38Diabetes with glycemic, unspecified, or no complications
0.166same as HCC 37
HCC 328Chronic kidney disease, moderate (stage 3B)
0.12725% of stage 4
HCC 329Chronic kidney disease, moderate (stage 3, except 3B)
0.12725% of stage 4
00.150.300.450.514
Highest in the examplePayment HCCSuppressed if stage 4 is documented
Hierarchies and interactions
Hierarchy in the kidney family: documented versus what pays [3]
HCC 326 (stage 5) trumps 327 (stage 4), which trumps 328 and 329. A member documented with both stage 4 and stage 3 scores 0.514, not 0.641.
0.514
pays
Stage 4 documentedHCC 327
0.127
suppressed
Stage 3 documentedHCC 329
0.641
wrong
Naive sumnot how the model works
PaysDropped by hierarchyCommon mistake
Normalization and the coding pattern adjustment
Divide by 1.079
Normalization, CY 2027
- Four-year regression on 2022 to 2025 FFS risk scores [2]
- Applies to the whole score, demographics included
- Implies 1.6% FFS risk score growth per year
Reduce by 5.90%
Coding pattern adjustment
- Statutory minimum, uniform across plans; CMS declined a tiered approach for 2027 [2]
What is left pays
Payment risk score
- Monthly payment is the adjusted score times the county rate; every 0.01 is a fixed share of it [4]
72-year-old woman, three documented conditions: 0.395 to 1.503
Community, not dual eligible, entitled by age; CKD stage 4, heart failure, type 2 diabetes with a chronic complication, all documented at face-to-face 2026 visits [3] [2]. Rounded to three decimals; CMS adds components such as frailty where applicable.
| Component | Source | Value | Running total |
|---|---|---|---|
| Female, 70 to 74 | Demographic factor | 0.395 | 0.395 |
| HCC 327, CKD stage 4 | Pays; HCC 329 would be suppressed | +0.514 | 0.909 |
| HCC 226, heart failure | Payment HCC | +0.360 | 1.269 |
| HCC 37, diabetes with chronic complications | Payment HCC | +0.166 | 1.435 |
| Diabetes x heart failure | Interaction term | +0.112 | 1.547 |
| Heart failure x kidney | Interaction term | +0.176 | 1.723 |
| Payment HCC count | Three HCCs, no count term | 0 | 1.723 raw |
| Divide by 1.079 | CY 2027 normalization | 1.723 / 1.079 | 1.597 |
| Reduce by 5.90 percent | Coding pattern adjustment | x 0.941 | 1.503 payment score |
Remove the kidney documentation
Without HCC 327 and the heart failure x kidney term the raw score falls by 0.690; one unstated stage moved the score by roughly 40 percent for this member [3].
1.723
Raw, stage 4 documented
1.033
-40%
Raw, stage not stated
1.503
Payment, stage 4 documented
0.901
-40%
Payment, stage not statedabout
RawPayment scoreWithout the stage
Takeaway
The RAF score is arithmetic on the note. The stage, the complication and the interaction are either written down at a 2026 face-to-face visit or they do not exist in 2027.
Which diagnoses enter the arithmetic
The operational side is in our 2027 unlinked chart review playbook; the model decision behind the 1.079 factor is in what moved in the 2027 model.
Counts
Does not
- A diagnosis that reached the plan only on an unlinked chart review record (except for enrollees who switched MA parent organization), or only from an audio-only call with modifier 93 or FQ [2]
Resets
- Every chronic condition, on January 1. The 2027 score is built only from 2026 encounters
Running it on your panel
Re-score a sample on the 2027 factorsOur free RAF score calculator applies the 2024 CMS-HCC coefficients and hierarchies to a member or a panel.
Check stage, type and complication languageThe HCC V28 coding reference lists the documentation each category requires.
Find documented-but-uncoded conditionsOur HCC risk adjustment AI maps de-identified charts through a 7,903-code crosswalk to the 115 payment HCCs, audited at 1,500 of 1,500 codes against CMS weights.
Keep a certified coder on every codeThe engine suggests; the coder confirms or rejects before anything goes anywhere.
Frequently asked questions
What is a RAF score?
A Risk Adjustment Factor score is the number CMS uses to scale a Medicare Advantage plan's payment for an enrollee. It combines demographic factors with coefficients for the hierarchical condition categories documented at eligible encounters in the prior year, then applies normalization and the coding pattern adjustment.
How is a RAF score calculated for 2027?
Add the demographic factor for the enrollee's segment to the coefficients of the payment HCCs documented from 2026 encounters, applying hierarchies so only the most severe category in a family pays and adding interaction terms. Divide by the 1.079 normalization factor and reduce by the 5.90 percent coding pattern adjustment.
Does a higher RAF score always mean more conditions?
No. The same conditions can produce different scores depending on segment, specificity and hierarchy. Stage 4 kidney disease carries 0.514 while stage 3 carries 0.127, and a condition documented only on an unlinked chart review or an audio-only call does not count at all for 2027.
Why does the normalization factor matter to a coder?
It does not change what to code, but it means a raw score of 1.723 pays as roughly 1.503 after normalization and the coding adjustment. Finance models and RAF targets need the 2027 factor of 1.079 to compare like with like.
- CMS fact sheet: 2027 Medicare Advantage and Part D Rate Announcement (April 6, 2026)
- CMS: Announcement of CY 2027 Medicare Advantage Capitation Rates and Part C and Part D Payment Policies (April 6, 2026)
- CMS: Announcement of CY 2024 Medicare Advantage Capitation Rates and Part C and Part D Payment Policies (March 31, 2023), Attachment VIII, Tables VIII-1 and VIII-4
- CMS: Payment Year 2024 MA Contract-Specific RADV Audit Methods and Instructions (August 28, 2026)
Checked October 25, 2026. Rules change; confirm against the source before relying on them. CMS has said the RADV audit schedule is subject to change.
Want the arithmetic run on your panel?
We re-score your members on the 2024 CMS-HCC model with the 2027 factors, show which conditions are documented but not coded, and flag codes the note does not support. A certified coder owns every code.