Risk Adjustment & Coding Optimization
Risk scores that reflect true patient complexity: accuracy, not inflation
Incomplete HCC documentation and coding produce RAF scores that understate patient complexity. We build accurate risk capture through structured HCC coding, prospective review, retrospective auditing and provider education, so every chronic condition that is present is assessed and documented at a real visit, to the standard a RADV reviewer expects.
- CMS-HCC V28
- Four-tier risk stratification
- Encounter-linked diagnoses only from CY 2027
Who it's for
Built for the organizations that carry this work
Each payment model uses a different risk model, but all of them now reward the same thing: conditions assessed and documented at a real visit.
Medicare Advantage groups
Capitated or percent-of-premium contracts paid on CMS-HCC V28 risk scores, where RADV recoupments can pass through to the practice.
MSSP ACOs
Benchmarks adjusted by risk score, with a positive-only cap on risk-score growth across the agreement period.
ACO REACH and LEAD participants
Benchmarks with a symmetric cap on risk-score growth, so coding cannot raise the benchmark beyond the cap and under-documentation cannot sink it.
IPAs, MSOs and FQHCs
Organizations that need one documentation standard across many practices and a recapture scorecard by site.
Commercial and Medicaid lines
HHS-HCC for marketplace plans and 3M CRGs for New York Medicaid managed care, where the same documentation discipline applies.
What we do
What you receive
Panel segmentation focuses resources where risk, and opportunity, is highest.
- 01
Four-tier risk stratification
Your panel segmented into Critical, Complex, Moderate and Low-risk tiers, so review effort and care management go where they matter most.
- 02
Prospective HCC chart review
Higher-risk patients reviewed ahead of their visits, with suspect conditions surfaced for the provider to evaluate and close at the encounter.
- 03
Provider coding education
Hierarchy-aware documentation for commonly missed categories: diabetes complications, CKD staging, CHF specificity, BMI, depression and morbid obesity, built on MEAT (Monitor, Evaluate, Assess, Treat).
- 04
AWV-based risk capture
The Annual Wellness Visit used as a high-yield touchpoint for HCC suspect review, with conditions assessed and documented at the visit.
- 05
Retrospective coding audits
Claims compared with EHR documentation to find missed or unsupported diagnoses, scored against RADV support standards and corrected at the source.
- 06
Provider-level RAF trend reporting
Ongoing monitoring and accountability, provider by provider, with a monthly recapture scorecard and coder-to-clinician feedback.
How it works
How an engagement runs
- 1
Baseline
Pull the plan's HCC gap report, claims, labs and pharmacy data, and measure current recapture by condition family.
- 2
Prioritize
Rank condition families and risk tiers by recapture gap, suspect volume and documentation risk.
- 3
Prepare the visit
Pre-visit suspect lists and a visit cadence for chronic patients, so nothing waits until December.
- 4
Document at the encounter
Template prompts and MEAT-standard notes; unconfirmed suspects are never coded.
- 5
Validate
Sample audits against RADV support standards, with corrections at the source before submission.
- 6
Monitor
A monthly recapture scorecard by practice and provider, and a record-retrieval process ready for audit requests.
Program facts
Codes, requirements and rules
What changed for 2027
As of September 9, 2026.
- CMS-HCC V28 stays fully phased in and was not recalibrated for 2027; the normalization factor is 1.079 (final).
- The MA coding pattern adjustment stays at the 5.90% statutory minimum (final).
- Diagnoses from unlinked chart reviews and audio-only encounters no longer count toward risk scores starting CY 2027, with an exception for members who switch MA organizations (final).
- In-home health risk assessment diagnoses are unchanged for now, but MedPAC has asked CMS to exclude them.
- CMS audits every MA contract annually with up to 200 records per contract and is working through payment years 2020 to 2025. Extrapolation was vacated by a federal court in September 2025; CMS's appeal is pending.
- The LEAD Model caps risk-score growth symmetrically at 3% (4% for High Needs) and phases in AI-inferred risk adjustment from 2028 to 2031.
Read the CY 2027 Medicare Advantage & Value-Based Care FAQ for the full set of changes and sources.
Risk models by payment program
Medicare Advantage
- Risk model
- CMS-HCC V28
- What to know
- Payment risk score = raw score ÷ normalization × (1 − coding-intensity adjustment). Unlinked chart-review diagnoses excluded from CY 2027.
Medicare Shared Savings Program
- Risk model
- CMS-HCC V28
- What to know
- Risk-score growth capped at +3% across the agreement period; uplift above the cap earns nothing.
ACO REACH (through 2026) and LEAD (from 2027)
- Risk model
- CMS-HCC V28-based
- What to know
- Symmetric 3% cap on risk-score growth; coding-intensity factor in REACH.
ACA marketplace
- Risk model
- HHS-HCC
- What to know
- Risk transfers between plans; reaches the practice only through a risk-sharing contract.
New York Medicaid managed care
- Risk model
- 3M Clinical Risk Groups
- What to know
- Reaches the practice only through value-based arrangements; weights are not public.
Where documentation most often falls short
Examples from the condition families we review. V28 payment HCC shown for reference.
Diabetes with complications (HCC 37)
- Common gap
- E11.9 coded when a complication is documented elsewhere
- What the note needs to show
- The complication named with its own specificity (CKD stage, neuropathy, retinopathy), status with the A1c value and date, and the plan.
Chronic kidney disease (HCC 326–329)
- Common gap
- N18.9 or stage 3 unspecified; a single low eGFR during an acute illness coded as CKD
- What the note needs to show
- The stage from the provider, two eGFR results at least 90 days apart with dates, the cause and the management.
Heart failure (HCC 226)
- Common gap
- "History of CHF" on a stable patient with no assessment
- What the note needs to show
- Type, acuity, ejection fraction with date, symptoms or volume status, and guideline-directed therapy.
Morbid obesity (HCC 48)
- Common gap
- A BMI code entered without an obesity diagnosis
- What the note needs to show
- The provider's diagnosis and class, the BMI value and date, the comorbidities it aggravates, and a plan. BMI alone never supports the code.
Major depressive disorder (HCC 155)
- Common gap
- Depression unspecified (F32.A) for a patient on treatment
- What the note needs to show
- Episode, current severity tied to the PHQ-9 and clinical description, remission status, restated each year.
Atrial fibrillation (HCC 238)
- Common gap
- "History of" code for a patient still anticoagulated for AFib
- What the note needs to show
- AFib type, current rhythm, rate or rhythm strategy, anticoagulation and the last ECG date.
FAQs
Common questions
What does the end of unlinked chart review diagnoses mean for my practice?
Beginning with 2027 payments, a diagnosis only counts if it is linked to a real encounter: yours, a specialist's or a hospital's. A plan can no longer add conditions from last year's records that were never addressed at a visit. If a patient with CKD stage 3 or heart failure was not seen and documented this year, that condition disappears from the risk score.
Am I exposed to RADV audits, or is that the plan's problem?
Now it is both. CMS audits every MA contract every year, sampling up to 200 records per contract. Auditors request the medical record from your office, an unsupported diagnosis is recouped from the plan, and many provider contracts pass that recoupment to the practice.
Why does every chronic condition need to be documented again each year?
HCCs reset on January 1. A condition counts toward the next payment year only if it is assessed and documented at an encounter during the year, so chronic conditions need to be reassessed at least every 12 months.
Can we code suspect conditions from labs or medications?
No. Labs, medications and prior reports are clinical indicators that prompt the provider to evaluate the patient. Only a condition the provider confirms and documents at a face-to-face visit is coded.
Do in-home health risk assessments still count?
For now, yes, but MedPAC has formally asked CMS to exclude them and to look at linked chart reviews next. The durable approach is that the treating providers assess and document conditions at real visits.
Read the full CY 2027 Medicare Advantage & Value-Based Care FAQ →
Ready to simplify the complex?
Let’s look at your program together.
A consultation is a working conversation about your panel, payers and EHR, and whether this service fits. Please do not send patient information.
Related services
CDI Review & Coding
Clinical documentation integrity reviews, coding accuracy audits and query workflows.
Learn morePractice Transformation & Quality Improvement
HEDIS and Star measures, care gap closure, AWV campaigns and care team redesign.
Learn moreACO & Value-Based Performance Enablement
Tie quality, risk and utilization work to Stars, withholds and shared savings.
Learn more