The MEAT Criteria Mistake That Almost Cost Us Everything

The audit letter arrived on a Tuesday. Twenty-three million dollars in questioned charges. The reason? We’d been teaching MEAT criteria coding completely wrong for three years.

We thought we needed all four elements: Monitor, Evaluate, Assess, AND Treat. So did our coders. So did our auditors. CMS only requires one. We’d been rejecting perfectly valid HCCs because they only showed three elements. Sometimes only two. Twenty-three million dollars of rejected accuracy.

The AND contra OR Disaster

Every training document said the same thing: “Document MEAT criteria.” Nobody specified whether that meant all four or any combination. We assumed all four because that seemed safest.

Williams monitored diabetes with quarterly labs. She evaluated the results. She didn’t document explicit assessment or treatment changes because the patient was stable. We rejected the HCC. CMS would have accepted it. Observing advancement alone is enough.

The cardiologist assessed CHF and treated it with medications but didn’t document observing advancement or evaluation in that specific note. Rejected. Should have been accepted. Assessment plus treatment equals valid MEAT.

Three years. Thousands of rejected codes. Millions in lost revenue. All because we misunderstood one word: “criteria” doesn’t mean “all criteria.”

The Over-Documentation Trap

Once we understood our mistake, we swung too far the other direction. If one MEAT element was enough, we’d document all four every time just to be safe!

The documentation evolved into robotic: “Monitored with labs, evaluated results showing stability, assessed as controlled, treated with continued medication.” Every condition. Every visit. Sounds compliant, right?

CMS flagged it as possible archetype abuse. The identical phrasing across hundreds of encounters looked suspicious. Over-documentation can cause audits just as fast as under-documentation. We’d traded one problem for another.

Real MEAT documentation should match clinical reality. Sometimes you monitor without treating (watching early CKD). Sometimes you treat without recent observing advancement (stable conditions). Forcing all four elements when they don’t naturally exist creates fiction, not compliance.

The Observing advancement Misunderstanding

“Observing advancement” doesn’t mean what we thought. We assumed it required labs, tests, or measurable observation. It doesn’t. Asking about symptoms counts. Reviewing functional status counts. Even observing appearance counts.

“Patient appears less short of breath” is observing advancement. “Continues to manage daily activities” is observing advancement. “Reports stable symptoms” is observing advancement. We’d been rejecting these as too vague. They’re actually perfect.

The flip side: ordering tests isn’t observing advancement unless you document reviewing results. We had hundreds of instances where providers ordered labs that showed kidney disease, but never documented looking at them. The lab result exists. The order exists. But without documented critique, observing advancement didn’t happen.

The Treatment Translation

Treatment confused everyone. Does “continue current medications” count? What about “no changes needed”? Is lifestyle counseling treatment?

Here’s what we learned the hard way: any active management counts as treatment. Continuing medications is treatment. Deciding not to change therapy is treatment (that’s clinical decision-making). Counseling is treatment. Even “watchful waiting” is treatment if you document it as a intentional approach.

But “stable” isn’t treatment. “No changes” isn’t treatment without setting. “Follow up in three months” isn’t treatment. Those are status updates, not active management documentation.

The psychiatrist who writes “continue Lexapro for depression” has documented treatment. The PCP who writes “depression stable” hasn’t. Same patient, same condition, completely different MEAT validity.

Your MEAT Reality Test

Pull twenty recent HCCs your team rejected for incomplete MEAT. Critique them looking for ANY single element done properly, not all four. How many would CMS accept that you rejected?

Check your archetype language. If every condition gets the same MEAT phrases, you’re creating audit targets. Real clinical documentation varies because real clinical care varies.

Count how many times providers document ordering tests without recording officially critique. Each one is failed observing advancement. The test results sitting in your EHR don’t count if nobody documents looking at them.

Look for “soft” observing advancement you’re missing: symptom discussions, functional assessments, visual observations. These count just as much as lab values but get overlooked all the time.

Critique your treatment documentation. Are you accepting only medication changes and procedures? You’re missing counseling, therapy decisions, and deliberate non-intervention strategies that absolutely count.

The truth about MEAT criteria is simpler than we made it. One element, properly documented, is enough. But that element needs to be real clinical care, not formulaic compliance language. Document what you actually do for patients. If you’re observing advancement, assessing the value of, assessing, or treating, say so in plain language. CMS isn’t looking for all four. They’re looking for proof you’re overseeing the condition. Any proof. One piece of proof. That’s all MEAT really requires.