MEAT Documentation: Proving Every Chronic Condition Annually

MEAT criteria in risk adjustment are the working test most coding teams use to decide whether a chronic condition in a note can be submitted: was it Monitored, Evaluated, Assessed or Treated at that encounter? It's a good test. It's also not the rule auditors apply.

CMS's RADV medical record reviewer guidance never mentions MEAT. The actual standard comes from the ICD-10-CM Official Guidelines and from CMS's documentation requirements. This article covers what MEAT means, what the official rules say, how to document the conditions that carry most of the risk adjustment weight under V28, and what never counts no matter how the note is written.

What MEAT stands for

Variants such as TAMPER (treatment, assessment, monitor/medicate, plan, evaluate, referral) do the same job with different letters. None of them is a federal standard.

Where MEAT comes from, and what the real rules say

MEAT is an industry convention. Health plans, coding vendors and compliance teams adopted it because it turns an abstract documentation requirement into something a coder can check line by line.

The official requirements sit in two places.

The ICD-10-CM Official Guidelines. For outpatient encounters, Section IV directs coders to report all documented conditions that coexist at the encounter and require or affect patient care, treatment or management. Conditions previously treated that no longer exist aren't coded as current. Uncertain diagnoses, such as "probable" or "rule out," aren't coded in the outpatient setting. These rules are in the FY 2027 guidelines, effective October 1, 2026.

CMS's RADV reviewer guidance. RADV reviewers require records that are clear and unambiguous, from a face-to-face encounter with an acceptable provider type, authenticated with a signature and credentials, and dated inside the data collection year. They submit conditions only to the level confirmed in the documentation.

MEAT is the practical bridge between those two. A condition with clear evidence of monitoring, evaluation, assessment or treatment almost always "affects patient care" under the guidelines. The reverse isn't guaranteed: a note can hit a MEAT letter and still fail RADV on a signature, a provider type or a date.

How many letters does a condition need?

Most internal policies treat one well-documented element as enough, as long as it clearly ties to the condition. Some plans and auditors expect more, especially for high-weight HCCs. Check your contract and your plan's documentation policy, and write the rule down. Coders shouldn't be deciding it chart by chart.

MEAT by condition: the four that matter most under V28

The Medicare Payment Advisory Commission found that after V28, the remaining MA coding-intensity gap is concentrated in diabetes, congestive heart failure, chronic obstructive pulmonary disease and morbid obesity. Those are the conditions most likely to be sampled, so they're where documentation has to be strongest.

Morbid obesity is the clearest example of a guideline coders miss. Under Section I.B.14, body mass index can be documented by other clinicians, but the associated diagnosis, such as obesity, has to come from the patient's provider. The BMI code is reported only as a secondary diagnosis.

Chronic kidney disease shows why stage matters. In a worked V28 calculation, documenting stage 4 instead of stage 5 changed the member's raw RAF score by 0.301. The stage has to be the clinician's statement, not the coder's reading of an eGFR.

The same condition, documented two ways

Illustrative note lines, not from a real patient.

Heart failure, weak:

PMH: CHF, HTN, DM2. Meds reviewed. Return in 3 months.

Nothing ties CHF to this visit. The medication review isn't linked to any condition. A reviewer can't tell whether heart failure was addressed or just carried forward.

Heart failure, strong:

Chronic systolic heart failure, EF 35% on echo 3/2026. Weight stable at 82 kg, no orthopnea, trace ankle edema. Continue carvedilol and furosemide; BMP today. Stable.

Every MEAT letter is present, and so is the type and acuity that determine the HCC. One line of clinical reasoning does more than a long problem list.

Diabetes with CKD, weak:

DM2. CKD. A1c 7.4.

The coder can't link the two conditions without the clinician's statement, and the CKD stage is missing.

Diabetes with CKD, strong:

Type 2 diabetes with diabetic CKD stage 3b; eGFR 38, stable from 41. A1c 7.4. Continue insulin glargine; review metformin dose given eGFR. Nephrology follow-up in May.

The link, the stage and the plan are all the clinician's words.

What never counts, however it's written

CMS's RADV reviewer guidance lists sources and documents reviewers reject. Training coders on this list prevents more audit failures than any acronym.

• Problem lists without evidence. Reviewers evaluate problem lists for whether conditions are chronic or past, and case by case when the list isn't clearly dated. A problem list on its own rarely carries an HCC.
• Pertinent negatives lists. Conditions listed as pertinent negatives aren't submitted.
• Patient-written lists and code lists. Lists of conditions written by the patient, or lists of code numbers without narrative, aren't acceptable.
• Health risk assessment screening results. Screening portions of an HRA aren't confirmed diagnoses.
• Referral and prescription forms. Conditions documented only on referral authorizations or drug order forms don't count on their own.
• Superbills. Not an acceptable source document.
• "Suggestive of" findings. Diagnostic test findings must be addressed by the treating or consulting provider to count.

Two ICD-10-CM rules add to the list: uncertain diagnoses ("probable," "suspected," "rule out") aren't coded in the outpatient setting, and conditions that were treated and no longer exist are coded as history, not as active disease.

Annual recapture: why the clock matters

The CMS-HCC model is prospective. Each year's RAF score uses diagnoses from qualifying encounters in the prior calendar year, and the slate resets every January. A condition the member has had for twenty years contributes nothing to next year's score unless it's documented, with support, in a qualifying encounter this year.

That applies to status conditions too. Amputation status, ostomy status and organ transplant status don't change, but they still need to appear in a qualifying encounter's assessment each year. A one-line exam finding or status statement in the assessment is usually enough.

Three practical consequences for a coding program:

1. Plan the visit, not the chart chase. A prospective list of conditions not yet documented this year, given to the clinician before an annual or chronic care visit, produces documentation that holds up. A retrospective search for diagnoses a year later doesn't, and from 2027 diagnoses submitted only through unlinked chart reviews no longer count toward risk scores.
2. Don't chase with addenda. Late addenda added months after a visit to support a diagnosis are exactly the pattern in the largest recent risk adjustment enforcement actions.
3. Track recapture by condition, not by HCC count. The goal is that every real chronic condition is documented properly once a year, not that the HCC total goes up.

Building a MEAT audit that matches RADV

Internal audits only protect you if they test what RADV tests. A simple per-HCC rubric works:

Score a random monthly sample of submitted HCCs, not just charts, with a second credentialed coder. Report the failure rate by check and by condition, and send the patterns back to coders and clinicians. If any HCC fails, delete it. A program that only ever adds is the profile RADV auditors look for first.

FAQ

Is MEAT required by CMS?

No. MEAT is an industry convention for showing that a condition was addressed. CMS's RADV reviewer guidance relies on the ICD-10-CM Official Guidelines and requires clear, authenticated documentation from an acceptable face-to-face encounter. MEAT is a useful way to check that standard, not the standard itself.

Does a condition need all four MEAT elements?

Usually not. Most internal policies accept one clear element tied to the condition, but some plans and auditors expect more for high-weight HCCs. Write your organization's rule down so coders apply it consistently.

Can a problem list support an HCC?

Rarely on its own. RADV reviewers evaluate problem lists for whether conditions are current and chronic, and an undated list is judged case by case. A condition needs evidence in the note that it was addressed at the encounter.

Does a medication list count as treatment?

Not by itself. A medication list doesn't show which condition the drug treats or that it was reviewed. Prescription order forms alone aren't acceptable to RADV reviewers. The note should tie the medication to the condition, for example "continue furosemide for chronic systolic heart failure."

How often do chronic conditions need to be documented?

At least once every calendar year, in a qualifying encounter. The CMS-HCC model resets annually, so a condition not documented during the year doesn't count toward the next year's risk score.

This article is general education on risk adjustment documentation. It isn't a coding determination for any specific patient or encounter.