V28 in 2026: the RAF points most groups are still leaving on the table.

Where we are in 2026The blend is gone. V28 stands alone now.

CMS moved to the V28 model on a three-year glide path. In payment year 2024 the risk score was a blend of 67 percent V24 and 33 percent V28. In 2025 that flipped to 33 percent V24 and 67 percent V28. For payment year 2026, the transition is complete: the risk score is 100 percent V28. The safety net that let a partial chart score partly on the older, more generous model is no longer there. Every point now runs through V28 alone.

That matters because V28 was not a cosmetic update. CMS remapped many conditions, retired a set of payment HCCs, and restructured the mappings so that some diagnoses that used to carry risk weight in V24 no longer map to a payment HCC at all. The net effect for many populations is a compressed RAF: the same patient, coded to the same real clinical picture, can score lower under V28 than under V24. The response is not to chase the lost points. It is to make sure the points the patient genuinely qualifies for are captured, documented, and defensible.

Under a blend, a thin chart still earned partial credit from the older model. In 2026 there is no partial credit. The chart either supports the HCC under V28, or it does not.

The reset nobody plans forEvery chronic HCC drops to zero on January 1.

Risk adjustment does not carry a diagnosis forward. HCCs are calculated from the diagnoses submitted during the calendar year. On January 1, the slate is wiped. A patient with diabetes with chronic complications, congestive heart failure, and chronic kidney disease had those conditions documented and coded last year. None of that carries into the new payment year. Each condition must be evaluated, documented, and coded again during the current year or it contributes nothing to the RAF. This is recapture, and it is the single largest source of preventable RAF leakage in most panels.

The leak is quiet because the patient is not new and the condition is not new. The staff know the patient has heart failure. The problem list says so. But a problem-list entry is not a coded, documented, face-to-face encounter for the current year. If the CHF is not addressed at a visit and captured on a claim this year, V28 scores it as if it were not there.

The three leaksWhere the points actually go.

Across most Medicare Advantage and risk-bearing panels, the same three failure modes account for the bulk of avoidable RAF loss. None of them is exotic. All of them are workflow problems, not coding-knowledge problems.

The third leak is the one that turns into a takeback. The first two lose points you never captured. The third loses points you did capture but cannot defend when the record is audited. A RADV or payer audit does not read the claim. It reads the note. If the note does not show MEAT, the HCC is unsupported and the payment tied to it is at risk. This is why the discipline is document, do not inflate. Every point you claim has to be a point you can prove.

Leak 2, in detailThe suspect workflow: evidence to confirmed diagnosis.

A suspect condition is a diagnosis the patient likely has, based on clinical evidence already in the record, that has not been formally documented and coded this year. The evidence is real: an A1c consistent with diabetes, an eGFR consistent with a CKD stage, a medication that treats a condition never listed as an active diagnosis, a prior-year HCC not yet recaptured. A suspect is not a guess and it is not a prompt to add a code. It is a prompt to have the clinician evaluate whether the condition is truly present, and if it is, to document and code it properly. If the evaluation shows the condition is not present, the suspect is closed and nothing is coded. That is the compliance guardrail.

The value of a disciplined suspect workflow is that it lifts the recapture and capture rate for conditions the patient genuinely has, while creating an auditable trail for the ones that were correctly ruled out. That trail is what separates legitimate risk adjustment from the practices CMS and the OIG have targeted. The goal is a complete and accurate picture of the patient, not a maximized one.

The recapture calendarRecapture is a year-long operation, not a Q4 scramble.

The groups that leak the least treat recapture as a rolling workflow that starts on January 1 and runs to December 31. The Annual Wellness Visit is the anchor. A well-run AWV is the single best vehicle for recapture because it is a comprehensive, preventive encounter where the full problem list is reviewed, every chronic condition can be assessed, and MEAT can be documented for each. Booking the AWV early in the year, and using it to sweep chronic conditions, front-loads the recapture instead of leaving it to a December panic.

MEATThe documentation standard that survives an audit.

MEAT is the four-part test that determines whether a documented diagnosis is supported: was the condition Monitored, Evaluated, Assessed, or Treated at the encounter. A code on a claim with a note that does no more than restate the diagnosis on the problem list does not meet MEAT. An auditor reading that note sees a diagnosis carried forward without evidence of active management, and the HCC is unsupported. MEAT is not a coding trick. It is the ordinary substance of a real clinical encounter, written down.

Point-of-care prompts are the mechanism that turns MEAT from a training slide into a habit. A prompt that surfaces an un-recaptured chronic HCC or an open suspect at the moment the clinician has the chart open, with the supporting evidence attached, is doing two things at once. It raises the recapture rate, and it does so at the only point in the workflow where the condition can be legitimately assessed and documented with MEAT. The prompt does not add a code. It hands the clinician the evidence and lets clinical judgment decide.

The deeper mechanics of this, including the reset report, the AWV-anchored recapture calendar, and the point-of-care prompt design, are laid out in the V28 RAF Optimization Playbook. For a worked example of what recapture and MEAT discipline look like applied to a real payment year, see the HCC V28 RAF Recapture case study.