Stroke Sequelae in V28: Which I69 Codes Carry an HCC, and Which Capture Nothing

"Sequela of stroke" is one of the quietest ways to lose an HCC. The acute stroke code (I63) is the one auditors talk about, but the money that goes missing without anyone noticing is in the I69 sequela family. Most of the codes a coder reaches for after a stroke carry no risk weight at all. A note can document a real, disabling deficit and still land on a code worth zero, because the deficit that got named was not a paralytic one.

This is the stroke deep-dive. Our acute vs history-of overview runs the acute-versus-resolved call across six conditions (stroke, MI, DVT and PE, respiratory failure, pressure ulcer, and AKI versus CKD) and gives the headline stroke rule: I63 for the event, Z86.73 for a history with nothing left behind, I69.35- for a documented deficit. This piece sits under that headline and does the one thing the overview did not have room for. It maps the whole I69 family, shows which of its codes actually capture in V28, and explains why the non-paralytic sequelae capture nothing even when the deficit is severe.

*Current as of July 2026. CMS-HCC model V28, payment year 2026 (V28 is 100% of the risk score this year). Reflects the FY2026 ICD-10-CM Official Guidelines, Section I.C.9.d, and the OIG acute stroke report A-02-23-01020. Verify every code and mapping against current CMS sources and your payer policy before you rely on it.*

The Acute Side, in One Paragraph

The overview handled the acute call, so this is the short version. An acute cerebral infarction, I63.-, maps to HCC 249, Ischemic or Unspecified Stroke, community RAF 0.239. Worth adding here, because the overview led with I63 alone: acute nontraumatic hemorrhage is a separate capture. Nontraumatic subarachnoid hemorrhage (I60.-) and nontraumatic intracerebral hemorrhage (I61.-) both map to HCC 248, Intracranial Hemorrhage, at the same 0.239 community weight. All three are in-the-moment codes. They belong on the encounter where the event is being worked up or actively managed, not on a stable follow-up months later. Everything past that acute window lands in one of two places: history (Z86.73) or a sequela (I69). The rest of this piece is about the second one, because it is where the code selection gets subtle.

The I69 Family, Decoded

An I69 code is built from three decisions, and only the middle one decides whether an HCC attaches. Read the code as three slots:

I69 . [cause] [deficit] [laterality]

The deficit digit is the whole game:

Read that as a rule. Only deficit digits 3, 4, 5, and 6 carry an HCC. Digit 5, hemiplegia and hemiparesis, maps to HCC 253, Hemiplegia/Hemiparesis. Digits 3, 4, and 6, the two monoplegias and other paralytic syndrome, map to HCC 254, Monoplegia and Other Paralytic Syndromes. Every other deficit digit maps to nothing, including the unspecified sequela codes (I69.30 and I69.90 and their siblings under every cause).

The cause digit does not change the HCC. Hemiplegia following cerebral infarction (I69.35-) and hemiplegia following subarachnoid hemorrhage (I69.05-) both land on HCC 253. So do the hemiplegia families under intracerebral hemorrhage (I69.15-), other intracranial hemorrhage (I69.25-), other cerebrovascular disease (I69.85-), and unspecified cerebrovascular disease (I69.95-). Six causes, one HCC, because the deficit is the same. The deficit digit is what moves the risk weight, not the cause.

The sixth character, laterality, does not change whether the HCC applies either. I69.351 (right dominant side) and I69.359 (unspecified side) both carry HCC 253. What laterality changes is audit quality. An unspecified-side sequela is a fully valid HCC capture, but it is also the kind of soft, under-documented code a reviewer circles. Name the side and the dominance when the record supports it. Section I.C.9.d.1 gives the default when the side is documented but dominance is not: right side defaults to dominant, left side defaults to non-dominant, and an ambidextrous patient defaults to dominant.

Real Deficits That Still Capture Nothing

Here is the part that costs coders, and it is the reason this deep-dive exists. The non-paralytic sequelae are not minor. Post-stroke aphasia (I69.320), dysphagia (I69.391), the cognitive deficits (I69.310 and its family), apraxia (I69.390), ataxia (I69.393), and facial weakness (I69.392) are all documented, managed, and clinically real. None of them carries an HCC on its own.

They are still worth coding, for an honest clinical picture and for other quality programs. They just do not risk-adjust. If a stroke patient has both aphasia and hemiparesis, the hemiparesis (I69.35-) is what captures, and the aphasia rides along at zero. If the note documents only the aphasia, the encounter captures nothing from the stroke, even though the patient is plainly impaired.

So the most useful sentence in this whole piece is narrower than it looks: "sequela of stroke, with a deficit" is not enough to capture. The deficit has to be a paralytic one (hemiplegia, hemiparesis, monoplegia, or another paralytic syndrome), and it has to be named as such in the note. A note that says "residual deficits from CVA" and stops there resolves to I69.30, which is worth zero.

TIA and History: the Honest Zeros

Two more codes belong on this map precisely because they are supposed to carry nothing.

When a stroke patient has no current deficit, Z86.73 is the accurate code, not a downgrade. The two ways coders go wrong here are symmetric. Reach for Z86.73 when a paralytic deficit actually persists, and you drop a real HCC 253 or 254. Reach past it for I63 on a stable follow-up with no acute event, and you have the overcode auditors pull.

Why OIG Made Acute Stroke the Anchor

In report A-02-23-01020, posted in June 2026, OIG estimated that CMS potentially overpaid Medicare Advantage organizations about $462 million for payment year 2021 based on acute stroke diagnosis codes that were not supported by the medical records. For all 97 sampled enrollees, OIG found the acute stroke codes the plans submitted were not supported by the records tied to those diagnoses. Our news desk has the dated write-up on the OIG stroke overpayment audit.

Keep a few things straight about that number. The $462 million is an OIG estimate, extrapolated from the sample, not a billed or recovered amount. The report covers multiple MA organizations, not one named plan, so no single plan owes $462 million. OIG did not recommend that CMS recover the estimated overpayment; it recommended CMS put a procedure in place to prevent this kind of overpayment going forward. And OIG's finding is that the diagnoses were unsupported in the record, which is a documentation and code-selection problem, not a finding of fraud.

For a coder, the report points straight at the residual-deficit documentation. The exposure is the gap between an acute code and what the record actually supports, and for a stroke patient months out from the event, what the record usually supports is a sequela or a history, not an acute infarction. The sequela only holds when the residual deficit is documented, and it only captures an HCC when that deficit is a paralytic one that is named. So the documentation that defends the code is exactly the thing that is easiest to leave vague: the specific, persisting, paralytic deficit, tied to the prior stroke, addressed at the visit.

The Overlap the Guidelines Allow

One legitimate combination gets undercounted. Section I.C.9.d.2 allows a code from category I69 alongside a current I60 through I67 code when the patient has a current cerebrovascular event and a deficit from an old one. A patient can present with a new acute infarction (I63.-) and carry residual hemiplegia from a prior stroke (I69.35-) on the same encounter. That is two codes for two distinct, documented facts, both supported, not a duplicate. Section I.C.9.d.3 draws the other boundary just as plainly: codes from category I69 should not be assigned if the patient does not have neurologic deficits. No deficit means no I69. It means history (Z86.73), which carries no HCC, correctly, because there is no ongoing condition to risk-adjust.

What the Note Has to Show

Work backward from what each code needs on the date of service.

Match the code to the clinical moment the note documents, not to the patient's stroke history in the abstract.

A Quick Self-Check on a Stroke Sequela

Before you finalize an I69 code, run it through four questions:

1. Does the note document a deficit that actually persists, or just "history of stroke"?

2. Is the deficit a paralytic one (hemiplegia, hemiparesis, monoplegia, other paralytic syndrome), which is the only kind that captures? Or is it cognitive, speech, or dysphagia, which do not?

3. If it is paralytic, is the side documented, with dominance where the record supports it?

4. If nothing persists, are you on Z86.73 instead of an I69 code that the record cannot support?

If the deficit on the chart is aphasia or "sequela of stroke" and you expected HCC 253 or 254, that is the gap this deep-dive is about. The paralytic deficit has to be named for the sequela to carry.

Where HCC Buddy Fits

None of this requires a tool, and none of it can be outsourced to one. What speeds it up is having the mapping in front of you while you read the note. The HCC Buddy encoder shows whether a code is billable and which HCC it maps to in V28, so you can see at a glance that I69.351 carries HCC 253, that I69.320 (aphasia) carries nothing, and that an unspecified I69.30 will not capture the way you expected. Check the deficit digit when it is not obvious, then make the call from the chart. The judgment stays yours.

For the documentation side, MEAT criteria covers what "addressed at the encounter" has to look like, and the RADV audit prep guide covers getting records ready before a letter arrives.