
MEAT Criteria: How to Document Diagnoses That Hold Up
MEAT criteria give documentation and coding teams a shared test for whether a diagnosis was actually managed at the encounter. Learn what each element asks for, which coding rule sits behind the mnemonic, and how to build a workflow that keeps risk-adjusted codes supported.
Every diagnosis in a patient's note has to earn its place on the claim. MEAT criteria are the quick check coders use to confirm the condition was monitored, evaluated, assessed, or treated during the visit.
A condition that only shows up in the problem list or past history gives the coder nothing to point to when a reviewer asks for support.
In this article, we cover what each element asks for, the coding rule behind the mnemonic, the documentation patterns that fail, and how documentation, coding, and audit teams can share the work.
TL;DR: MEAT Criteria at a Glance
In short, MEAT criteria come down to five points:
What MEAT means: Monitor, Evaluate, Assess, or Address, and Treat.
Where it comes from: An industry convention for applying the ICD-10-CM Official Guidelines.
Why it matters: Risk adjustment reviewers look for this evidence on every diagnosis.
How much is enough: One clearly documented element per diagnosis.
Who owns it: Clinicians document, coders confirm, and auditors test.
What Are MEAT Criteria in Medical Coding?
MEAT criteria are a documentation test that coders, clinical documentation integrity (CDI) teams, and auditors use to decide whether a diagnosis in the note is supported. The acronym stands for Monitor, Evaluate, Assess, Address, and Treat.
AAPC, the coding credentialing organization, breaks each letter into concrete evidence in its guidance on MEAT in risk adjustment documentation. The same guidance adds that a simple list of diagnoses does not meet the definition of an assessment and plan.
MEAT itself appears in no federal regulation. The convention exists to help teams apply the current ICD-10-CM Official Guidelines, which instruct coders in outpatient settings to code every documented condition that coexists at the encounter and requires or affects patient care, treatment, or management.
The same guidelines base code assignment on the provider's diagnostic statement. MEAT supports that rule by checking whether the note shows the condition was managed at this visit, leaving clinical judgment about the diagnosis with the provider.

MEAT Criteria Elements and What Each One Requires
Each letter describes a different kind of evidence, and any one of them can carry a condition when it is specific to the encounter. The examples below are illustrative note language, written the way a reviewer would hope to find it.
Below, we break down what each MEAT element looks for in the note, with an example of each:
Monitor
Monitoring shows the clinician is tracking a condition's signs, symptoms, progression, or regression. A line such as "atrial fibrillation, rate controlled on current regimen, no palpitations reported since last visit" ties the condition to this encounter.
Status words work well when you document the observation behind them. "Stable" paired with what the clinician checked gives a reviewer something to confirm.
Evaluate
Evaluation shows the clinician reviewed test results, medication effectiveness, or response to treatment. "Major depressive disorder, recurrent, moderate, PHQ-9 reviewed today and improved from the prior visit" is a clear example.
The review has to connect to the condition. A lab value pasted into the note without comment shows the result exists and leaves the reviewer to guess whether anyone looked at it.
Assess or Address
Assessment shows the clinician judged the condition's status or took a step to work it up, such as ordering tests, reviewing records, discussing options, or counseling. "Chronic obstructive pulmonary disease (COPD) with increasing dyspnea on exertion, pulmonary rehabilitation discussed, spirometry ordered" for example, would meet this element.
Some organizations teach the letter as Assess and others as Address. Both versions point to the same evidence: a documented clinical judgment about the condition at this visit.
Treat
Treatment shows a medication, therapy, procedure, or referral tied to the condition. "Rheumatoid arthritis, continue methotrexate, refill sent, rheumatology follow-up scheduled," for example, would meet it.
The plan carries the weight when it names the condition each treatment is for, which a medication list on its own rarely does.
What Does a MEAT-Supported Note Look Like?
Consider a single follow-up visit with four chronic conditions on the chart. A reviewer reads each condition separately, so one strong entry doesn't support the others.
The table below shows how a reviewer would read the note for each of the four conditions:
Condition on the chart | What the note shows | MEAT element present | Reviewer's read |
|---|---|---|---|
Heart failure with reduced ejection fraction | Weight and ankle edema checked, diuretic dose adjusted | Monitor, Treat | Supported |
Peripheral artery disease | Ankle-brachial index result reviewed, walking program discussed | Evaluate, Assess | Supported |
Obesity | Body mass index (BMI) recorded in nursing vitals, no provider mention | None from the provider | Not supported as written |
Chronic kidney disease, stage 3 | Listed in the problem list only | None | Not supported as written |
The obesity row shows a detail many teams miss. The same ICD-10-CM guidelines allow the BMI to come from other clinicians' documentation, and they require the associated diagnosis, such as obesity, to come from the patient's provider.
A plan line for the last two conditions would change the result only if the provider actually addressed them at the visit. MEAT documents care that happened, and it never justifies adding diagnoses the visit did not address.

Why Do MEAT Criteria Matter for Risk Adjustment?
In risk-adjusted programs such as Medicare Advantage, documented diagnoses map to Hierarchical Condition Categories (HCCs) that feed a patient's risk score and the plan's payment. The mechanics of HCC medical coding make each supported diagnosis matter to both payment and audit exposure.
The following two developments raise the stakes for documenting MEAT evidence at the encounter:
Oversight Focuses on Diagnoses With No Documented Care
The Office of Inspector General (OIG) at the U.S. Department of Health and Human Services found that diagnoses reported only on health risk assessments (HRAs) and HRA-linked chart reviews drove an estimated $7.5 billion in Medicare Advantage risk-adjusted payments in a single payment year. The OIG report on health risk assessments noted that 1.7 million enrollees had no follow-up visits, procedures, tests, or supplies for those diagnoses in the encounter data.
The finding gives coding and compliance teams a clear test. Each submitted diagnosis should trace to an encounter where the note shows the condition was managed.
CMS Now Excludes Diagnoses That Lack a Linked Encounter
The Centers for Medicare and Medicaid Services (CMS) has finalized policies that exclude diagnoses from audio-only encounters and from chart review records not linked to a specific beneficiary encounter when it calculates Medicare Advantage risk scores. The CMS rate announcement fact sheet notes one exception for unlinked chart reviews, covering beneficiaries who move from one Medicare Advantage organization to another.
Retrospective review can still surface gaps, and the evidence now has to sit in a real encounter for the diagnosis to count. Teams building out defensible risk adjustment coding can treat MEAT written at the point of care as the daily checkpoint that keeps both rules in view.
What Documentation Fails MEAT Criteria?
Many MEAT failures come from habits that save time in the moment and cost support later. The patterns below show up across specialties and in both inpatient and outpatient notes.
The table below lists common failure patterns, why each one falls short, and what supports the diagnosis instead:
Pattern in the note | Why it falls short | What supports the diagnosis |
|---|---|---|
Problem list entry with no mention in the assessment or plan | Shows the condition exists and says nothing about this visit | A status, result, or plan entry for the condition in the assessment |
"History of" language for a condition still under treatment | Reads as resolved, so the current condition goes unsupported | Current-tense status that matches what the clinician is managing |
Assessment copied forward across several visits | Gives a reviewer no evidence the condition was looked at again | Updated findings, results, or plan changes from this encounter |
Lab result or medication list with no link to a diagnosis | Leaves the reviewer to infer which condition it supports | A sentence tying the result or medication to the named condition |
Diagnosis captured only on an HRA or chart review | May have no linked record of care behind it | Documentation of the condition in an encounter note |
Unspecified diagnosis where the note supports more detail | Can map to a different category than the condition warrants | Type, stage, severity, or manifestation stated in the note |
Each fix takes seconds at the moment of documentation. That spares your team provider queries, coder time, and the data corrections recovery support required after submission.
Most of these gaps start at capture, which is where our documentation work begins. At DeliverHealth, we provide AI-powered clinical documentation and coding solutions, and InstaNote turns ambient conversation or dictation into a structured note for inpatient and outpatient care.
InstaNote also offers coding suggestions based on the note, which help your providers hand off a more complete note before sign-off. Your coders still review and approve the final codes.
InstaNote's speech recognition is designed not to add anything that wasn't said, and note templates follow your organization's and each provider's requests, so the status, results, and plan behind each condition land where reviewers look for them. InstaNote is trusted by more than 40,000 healthcare providers.
Book an InstaNote demo to watch a sample note in your specialty move from capture to coding suggestions.
How to Build MEAT Into Your Documentation and Coding Workflow
MEAT holds up when each team owns a clear part of it. Clinicians create the evidence, coders confirm it, and auditors test whether the process works over time.
You can keep MEAT evidence in place from the encounter to submission with the following three checkpoints:
1. At the Point of Care
Give your clinicians templates that prompt a status, a result, or a plan for each chronic condition they address. Short prompts in the assessment and plan section tend to change habits faster than retraining because they fit how notes are already written.
AI documentation helps here when it stays faithful to what was said in the room.
As our Director of Product Implementation, Garrett Dunne, puts it, "There's not a situation where the engine is being essentially, the model is being asked to infer a diagnosis. We wouldn't allow it to do that."
For organizations that want a review step before sign-off, Clinical Documentation Review in InstaNote (CDR) reads the draft note for completeness, consistency, specificity, and quality support. Established frameworks back its checks, including HCC risk adjustment, MEAT and TAMPER documentation criteria, and Healthcare Effectiveness Data and Information Set (HEDIS) quality measures.
CDR surfaces gaps for the clinician to review. The clinician decides on each item and retains final content and sign-off.
2. In the Coding Review
Coders confirm that the note meets MEAT criteria for each risk-adjusting diagnosis before submission and query the provider when a gap remains. A consistent query template keeps those questions neutral and specific to the documentation.
Where AI-assisted coding is part of the workflow, InstaCode works from any documentation source, including InstaNote, dictation, and existing electronic health record (EHR) notes. Your coders, our AAPC- or AHIMA-credentialed coders, or both stay in the loop.
Every coded encounter stays auditable. InstaCode tracks the AI's output, human changes, the reasoning behind them, and any further edits by human auditors.
3. In Internal Audits
Sample risk-adjusted diagnoses on a regular schedule and score each one against the four MEAT criteria elements. Track failures by pattern, provider group, and specialty so education goes where the gaps are.
Close the loop by sharing findings with the clinicians and coders involved. A coding compliance program that records each correction gives you a documented account of how your organization responds to what it finds.

Frequently Asked Questions (FAQs)
Below are answers to the questions coding, CDI, and compliance teams most often ask about MEAT criteria:
Does a Diagnosis Need All Four MEAT Elements?
No. Most coding and audit programs treat one clearly documented element as enough to support a diagnosis at the encounter.
More elements make the support easier to defend, and individual health plans may set stricter expectations, so check the policies that apply to your contracts.
What Is the Difference Between MEAT and TAMPER?
TAMPER is a related industry convention, commonly expanded as Treatment, Assessment, Monitor or Medicate, Plan, Evaluate, and Referral. Many teams treat it as the stricter of the two.
Both are industry conventions that sit outside CMS regulation. Each aims for the same result: a note showing the condition was managed at the visit.
Can a Problem List Entry Meet MEAT Criteria?
A problem list entry on its own shows that a condition exists in the record. Support comes from the assessment and plan, where the clinician records the condition's status, results, or treatment at the current visit.
Who Is Responsible for MEAT Documentation?
The provider who documents the encounter creates the evidence, since code assignment rests on the provider's documentation.
Coders, CDI specialists, and auditors check that evidence and raise queries when it falls short, so accountability runs across the whole workflow.
Can AI Check Notes Against MEAT Criteria?
AI can flag diagnoses that lack a status, result, or plan, and it can do so before the note is signed.
Your clinicians still decide what the note says, and your coders still approve the final codes, so the human review step stays in place.
Final Thoughts on Applying MEAT Criteria
MEAT lasts as a convention because it asks what every reviewer already wants to know, whether the note shows the care the code describes. Teams that build the answer into templates, coding review, and audits spend far less time defending diagnoses after submission.
The clinical note is where the revenue cycle starts, and MEAT is one of the clearest ways to check that the start is sound. That checkpoint is also where our coding work picks up.
At DeliverHealth, our AI-powered solutions cover both halves of that work: InstaNote builds the note, and InstaCode turns it into accurate, defensible codes. InstaCode delivers 95% coding accuracy with human-in-the-loop review and typically returns billable codes within about 24 hours.
Schedule an InstaCode demo to see sample encounters in your specialty become audit-ready codes.
Stay Updated
Subscribe to our newsletter for the latest healthcare AI insights and company updates.
