Problem Lists Alone Do Not Validate HCCs: What RADV Auditors Check

Short answer

No. A problem list entry by itself won't hold up under RADV review.

CMS guidance says reviewers should evaluate problem-list conditions for

chronicity, consistency with the current encounter, and support in the full

medical record: history, medications, and final assessment included.

CMS also says patient-written lists aren't acceptable, and lists of code numbers

without narratives aren't acceptable. The problem list can point you toward a

diagnosis, but it doesn't do the chart-support work for you.

What CMS reviewer guidance actually says

The trap isn't that problem lists are worthless. The trap is that teams treat

them like they close the loop when they don't.

CMS RADV reviewer guidance says problem lists are evaluated case by case,

especially when a list isn't clearly dated as part of the face-to-face encounter

or when multiple dates appear around the date of service. A condition that lives

on the active list but shows no current discussion in the note is weak support,

full stop.

Three parts of the guidance matter most for coders doing daily QA.

CMS tells reviewers to check the full medical record

The guidance says reviewers should evaluate the problem list for whether

conditions are chronic or past, and whether they're consistent with the current

encounter. It also directs reviewers to check the full record, including history,

medications, and final assessment.

That means a copied-forward diagnosis with no active discussion is risky even

when it sits on the active problem list. The list records what the provider

knows. The note shows what was addressed at the DOS.

Some problem-list formats are plainly weak out of the gate

CMS says patient-written lists aren't acceptable. Lists of code numbers without

narratives aren't acceptable. If the only support you have is a structured table,

a pasted diagnosis panel, or a code-only grid with no clinician narrative, that's

already a documentation problem before you even get to MEAT.

No documented condition means don't submit the record

CMS says not to submit a medical record if no conditions are documented, and it

gives examples: visits showing only vitals, screening language, or a generic

follow-up note with no positive findings. This matters because a lot of

problem-list disputes are really current-note disputes. The condition may exist

deep in the chart history, but if the selected encounter doesn't document it,

the record is weak for RADV purposes.

See the MEAT criteria hub for the full documentation standard

and what "addressing a condition" looks like encounter by encounter.

Why problem lists are a recurring QA miss

Problem lists are designed to remember the chart, not to prove the encounter.

EHR problem lists carry diagnoses forward long after the clinical picture

changes. Cancer history can stay listed as active. Diabetes with CKD can stay on

the list even when the provider never addressed the renal disease at the current

visit. A condition can stay listed after the provider updated the specificity

somewhere else in the chart and nobody updated the list to match.

CMS flagged this pattern in its RADV Medical Record Checklist and Guidance.

The checklist says a diagnosis may never drop off the problem list even after the

patient is no longer suffering from that condition, and it warns that the problem

list may not even reflect the HCC the MA contract submitted for payment.

The OIG kept the same pressure on in its October 2024 report on Medicare

Advantage health risk assessments and HRA-linked chart reviews. That report found

that HRA-linked diagnoses drove an estimated $7.5 billion in MA risk-adjusted

payments for 2023, with unsupported diagnoses as a key driver. It's not only

about problem lists, but it reinforces the same operational point: unsupported

diagnosis sources create real payment risk.

What to verify before keeping the HCC

These are the questions that should gate your decision at the chart level, not

the problem list.

Is the diagnosis supported in the current encounter note?

Look for current provider language, not a carried-forward label. The most

defensible note shows the condition in the assessment, plan, or another clear

narrative section tied to that DOS. A condition mentioned only in "Active

Problems" with no corresponding text in the body of the note is not supported

for that encounter.

Does the record show the condition is still clinically relevant?

CMS reviewer guidance points back to the full record. Check whether the

diagnosis is consistent with history, the medication list, and the final

assessment. If the problem list says heart failure but the note, the med list,

and the plan are all silent on it, don't assume the condition was managed at that

visit.

Is the specificity still accurate?

Stale problem lists can preserve an older description after the provider has

documented something narrower, broader, or resolved elsewhere in the chart.

Recheck whether the current note actually supports the ICD-10-CM code you're

planning to keep. Use the ICD-10 encoder to compare the code path

against what the note says.

Is this tied to a face-to-face supported source?

RADV centers on the medical record selected for the relevant encounter. A

condition that only appears in a disconnected list, a referral form, or a

historical summary with no presence in the encounter note isn't face-to-face

encounter support.

What QA leads should check this week

Start with the pattern that causes the most misses: diagnoses that look familiar,

so nobody questions them.

Pull a sample of charts where the only visible support is the problem list

You're looking for charts where the code felt safe because it's been there

forever. Those are often the ones nobody reviews. Run a few, look at where the

current encounter actually supports the condition, and see how many hold up.

Separate list presence from note support in your QA scoring

Don't score a condition as "supported" because it appears on the active list.

Require the reviewer to identify where the current encounter supports it. If they

can't point to the note, the condition isn't supported at that encounter.

Re-educate on query timing

When the note is thin, fix it the right way. CMS reviewer guidance says

acceptable query responses need to become part of the official medical record,

signed and dated by the treating provider in a timely manner. A coder can't add

a diagnosis to the chart after the fact because the problem list suggests it.

Check the provider query templates guide for

compliant query patterns when the note needs clarification before you can keep

the code.

Cancer history, diabetes complications, and resolved conditions first

These are the most common stale-list patterns. If your QA team checks one thing

this week, check whether active-diagnosis wording in the problem list still

matches the current assessment and treatment story in the note. A year-old

"active" status that the note never addresses is a RADV miss waiting to happen.

See the RADV audit prep guide for the full

pre-submission chart review workflow.

Where HCC Buddy fits

HCC Buddy helps with the reference work. You can check the ICD-10-CM code path,

compare specificity options, verify the V28 HCC mapping, and move through

supporting references without switching tabs.

It doesn't turn a problem list into encounter support. You still have to read

the note, confirm the diagnosis is current and documented, and decide whether

the chart is defensible. That judgment call is yours.

Sources

CMS Medicare Advantage RADV program page

CMS RADV Questions and Answers, updated March 4, 2026

CMS RADV Medical Record Checklist and Guidance

CMS Contract-Level RADV Medical Record Reviewer Guidance

HHS OIG report on Medicare Advantage HRAs and HRA-linked chart reviews, posted October 24, 2024 (OEI-03-23-00380)