The 2021 revisions to the E/M office-visit guidelines changed how you pick a level for CPT codes 99202 through 99215: you now choose by medical decision making or by total time on the date of the encounter, not by counting history and exam bullets. An ambient AI scribe helps by drafting the note so the problems, data, and risk you actually discussed are written down, and by reflecting the time-relevant work of the visit, for you to verify. Patient Square hands that back about two minutes after the visit as a draft with ICD-10 suggestions. It does not guarantee coding compliance and it does not bill. You review, confirm the code, and sign.
Key takeaways
- The 2021 E/M guidelines let you code office visits (99202 to 99215) by MDM or by time.
- An ambient scribe drafts the note so MDM elements discussed in the room are documented.
- It reflects time-relevant activity, but you attest to the total time you actually spent.
- ICD-10 codes come back as clinician-verified suggestions, not final billing decisions.
- It never guarantees coding compliance and never bills; the signed record and code are yours.
Year the E/M office-visit guidelines were revised
From visit end to draft note in hand
Office-visit CPT codes the MDM-or-time rule covers
Source: CMS and AMA guidance on the 2021 E/M office-visit changes; see the sources below.
What the 2021 E/M guidelines actually require
For office and outpatient visits, CMS and the AMA revised the coding framework effective January 1, 2021. The core change: level selection is based on either medical decision making or total time on the date of the encounter. History and exam still belong in a competent note, but they no longer set the level. This applies to CPT 99202 through 99215.
Medical decision making, under the revised guidelines, rests on three elements: the number and complexity of problems addressed, the amount and complexity of data reviewed and analyzed, and the risk of complications from the visit and its management. Two of those three, at a given level, support that level.
Time, the alternative path, means total time spent on the date of the encounter, including non-face-to-face work like reviewing records and documenting the note, not just the minutes in the room. If you select by time, your documentation has to support the total time you are attesting to.
Both paths share a requirement that is easy to lose in the abstraction: the note has to actually reflect what happened. That is where documentation load and coding accuracy meet.
How an ambient scribe helps you capture MDM
MDM is a documentation problem before it is a coding problem. If the problems you addressed, the data you reviewed, and the risk you weighed never make it into the note, the visit looks simpler on paper than it was in the room, and it may be under-documented for the level of work you did.
An ambient scribe drafts the note from the conversation, so the problems discussed, the labs or records you talked through, and the risk you raised with the patient get written down instead of forgotten. Patient Square returns that draft about two minutes after the visit. You read it against the MDM framework, confirm the elements are captured accurately, and sign.
The scribe is not deciding your MDM level. It is making sure the raw material for that decision is documented in the note. Whether the documentation supports a given level is your call, checked against CMS and AMA guidance. That distinction is the whole point of the next section.
The line the scribe does not cross
Say this part clearly, because coding is where overpromising gets clinicians in trouble.
The scribe drafts a note and suggests ICD-10 codes. It does not guarantee coding compliance. It does not choose your E/M level for you. It does not bill anything and it has no connection to any claims system. The ICD-10 suggestions are clinician-verified: you confirm or change every one, and the diagnosis is yours.
The prescription draft follows the same rule. It reflects the plan you discussed in the visit; it does not screen for interactions, dosing, or contraindications, and it will not block or override anything. That review is yours to run, exactly as you run it today. It is a draft only. It never sends to a pharmacy. There is no e-prescribing. You prescribe through your own system.
None of this is a coding engine. It is documentation support that gives your coding process better raw material. The signed record, the E/M level, and the billing code stay with you and your compliance workflow.
MDM elements and where the note comes from
Here is how the pieces line up, so you know what to verify.
| E/M path | Element | Where the draft note reflects it | What you verify |
|---|---|---|---|
| MDM | Problems addressed | Problems discussed in the visit | Count and complexity match the encounter |
| MDM | Data reviewed | Labs, records, studies you talked through | Documentation supports what you reviewed |
| MDM | Risk | Risk and management options discussed | Risk level reflects your clinical judgment |
| Time | Total time | Time-relevant activities of the encounter | Your attestation of total time on the date |
| Diagnosis | ICD-10 | Suggested from what was said | Confirm or change every code |
The pattern holds across every row: the scribe drafts, you verify. It never converts documentation into a guaranteed code.
Time-based selection, and honest limits
If you code by time, the scribe can reflect the time-relevant work discussed in the visit, but it cannot know the total time you spent unless you tell it. Reviewing records before the patient arrived, documenting after they left, coordinating care later that day, those count toward total time on the date of service and they live partly outside the recorded encounter. Your time attestation is yours to make and document.
So treat any time figure in a draft as a starting point, not a total. The same caution applies to MDM: the scribe captures what was said, but it does not read your intent or your outside review unless it was part of the conversation. Verify against the actual visit.
When this matters for your practice
If Medicare and commercial payers make up a real share of your panel, under-documented visits are lost work and over-documented ones are a compliance risk. Better documentation of MDM and time protects you in both directions, and an ambient scribe raises the floor on documentation without touching the coding decision.
For adjacent workflows, how ICD-10 suggestions work explains the code-suggestion side, reducing claim denials through documentation covers the downstream billing angle, and SOAP note quality goes into the note itself. On the trust and compliance side, the HIPAA, BAA, and consent page has the specifics.
Nothing here should be read as coding advice for your specific situation. The 2021 guidelines are published by CMS and the AMA, linked below, and your coder or compliance lead is the authority on how they apply to your visits. Want to see the documentation half in practice?