Retrospective Review
Coders review past encounters before the submission deadline to find supported diagnoses that never reached claims and flag codes the record does not support.
Retrospective Review →
With CMS-HCC V28 fully phased in and RADV activity rising, your HCC coding vendor shapes both revenue accuracy and audit exposure. Here is what to evaluate, what to ask, and which red flags should end the conversation.
Three shifts have raised the bar. Any vendor you shortlist should explain how its process handles each one.
CMS phased in V28 over payment years 2024 and 2025, and payment year 2026 runs entirely on it. V28 restructured the HCC hierarchy and dropped many diagnosis codes from payment, so V24-era rules and mappings now mislead.
CMS has said it intends to audit far more Medicare Advantage contracts, and ACA issuers face HHS-RADV each year. Recovery rules have been litigated, but the core test is unchanged: every submitted HCC needs record support.
Federal oversight and enforcement have targeted diagnoses found only in chart reviews or health risk assessments, and programs that searched for adds without removing unsupported codes. Vendor method is now a compliance question.
Most programs need more than one review type. A capable vendor supports all three and explains where each fits in your year.
Coders review past encounters before the submission deadline to find supported diagnoses that never reached claims and flag codes the record does not support.
Retrospective Review →Pre-visit chart preparation that surfaces suspected or previously documented conditions so the provider can assess and document them at the next visit.
Prospective Review →Coding review after the visit and before the claim is final, so gaps can be fixed through a compliant provider query while the encounter is fresh.
Concurrent Review →No review beats the records behind it. Ask whether the vendor also retrieves charts from EMRs, fax, and onsite sources with secure chain of custody.
Chart Retrieval →Compare proposals side by side. Ask for written answers and sample output, not slides.
The vendor should report diagnoses to add and to delete, with both counts in every report. Deletes go to CMS as chart review delete records, cleaning your data before an audit does.
Each coded condition should cite its support under MEAT (Monitor, Evaluate, Assess or Address, Treat) or TAMPER, from an acceptable face-to-face or audio-video encounter with an acceptable provider type, within the year.
Ask what share of your assigned team holds the AAPC Certified Risk Adjustment Coder (CRC) credential, what else they carry (CPC, CCS), and how V28 retraining was handled.
Ask how accuracy is measured, what share of charts gets second-level QA, and whether high-impact HCCs get dual review. Tracking inter-rater reliability (IRR) shows consistency, not just averages.
Expect a signed BAA, HIPAA-compliant PHI handling, role-based access, guidelines tied to the ICD-10-CM Official Guidelines and Coding Clinic, and a query policy that never leads the provider.
Get turnaround in writing, tied to your deadlines and volumes. Per chart pricing suits retrospective projects; hourly or FTE pricing suits ongoing prospective and concurrent work. Tie pricing to accuracy, never to HCCs added.
Any one of these on the left is a reason to slow down. Two or more is usually a reason to walk away.
Send these to every shortlisted vendor. Strong vendors answer directly, with sample reports.
Valiant Lifecare works this way by default: credentialed coders, two-way findings, MEAT-linked evidence, and QA review, delivered from our Pune delivery centre with US-based account management. See how it fits together on our risk adjustment services and risk adjustment optimization pages, or read how we support plans on our health plans page.
Straight answers for health plan and provider group teams comparing vendors.
A risk adjustment coding company reviews medical records for Medicare Advantage, ACA, and other risk-based members, abstracts diagnoses that are supported by face-to-face documentation, maps them to HCCs, and returns coded results that a health plan or provider group can submit or correct. A good vendor also flags unsupported diagnoses for deletion and feeds documentation issues back to providers.
RADV audits check whether submitted diagnoses are supported by the medical record. If a vendor only looks for diagnoses to add, unsupported codes already in your data stay there and become audit exposure. Two-way coding reports both adds and deletes, so the data you submit reflects what the record actually supports.
CMS phased in the V28 model over payment years 2024 and 2025, and payment year 2026 is calculated fully on V28. V28 restructured the HCC hierarchy and removed a number of diagnosis codes from payment HCCs. Your vendor's coding guidelines, software mappings, and coder training all need to reflect V28, not the older V24 model.
Look for coders who hold the AAPC Certified Risk Adjustment Coder (CRC) credential, often alongside a CPC or an AHIMA credential such as CCS. Ask what share of the team assigned to your work holds the CRC, how often coders are retrained, and who reviews their work.
The most common models are per chart, per hour, and dedicated FTE. Per chart pricing is predictable for retrospective projects but should be tied to accuracy standards. Hourly and FTE models fit ongoing prospective or concurrent programs. Be cautious of pricing tied to the number of HCCs added, since it rewards volume over accuracy.
Yes. Our free HEDIS and risk adjustment gap review looks at a sample of your records and coding output and shows where documentation gaps, unsupported codes, and missed conditions sit, so you can judge our work before signing a larger engagement.
Send a sample of records and coding output. We will show where supported conditions are missing, where codes lack support, and what a two-way program would look like.