Office Visit CPT Codes 99202-99215 Explained (2026)

Knowing the right CPT code ranges is the easy part. The revenue practices lose comes from documentation gaps no alert ever surfaces, and fixing that starts here.

Every practice needs two number ranges committed to memory: 99202-99205 for new patients and 99211-99215 for established patients. Those ten codes cover virtually every outpatient office visit a practice bills. But knowing those numbers is only the entry requirement.

The harder problem sits one layer deeper, in whether the clinical documentation captured at the encounter can actually defend the level that was billed. Most practice administrators and billing decision-makers believe that once coders know the correct CPT code ranges and time thresholds, E/M coding accuracy is essentially solved and revenue leakage is a residual, manageable exception. That assumption leaves a gap between what administrators believe and what the documentation reveals, and that gap is where most practices quietly lose money.

See our AI medical scribe for how this works in practice.

"Patients are being billed for CPT codes (e.g. 90838 add-on to 99204) they believe are unjustified, highlighting confusion and potential misuse of office visit and psychotherapy CPT code combinations."

— what we hear from medical billing professionals

Most practice administrators treat the code list as the finish line. Once the billing team can recite the ranges, the assumption is that the revenue problem is solved. It is not. According to JNC Healthcare Compliance Group (2022), approximately 55% of E/M claims contain coding errors, and the root cause is almost never code ignorance. The gap is documentation quality. That distinction is what separates practices that capture their full earned reimbursement from those that quietly leave money behind, encounter after encounter, without ever seeing the pattern.

55%

of E/M claims contain coding errors

CPT 99201 no longer exists. It was retired effective January 1, 2021, when the AMA overhauled the E/M guidelines, making 99202 the lowest billable code for new patients.

Practices still occasionally see 99201 surface in older templates or clearinghouse edits; that code will be rejected every time. The current code set breaks into two families. New patient codes run 99202 through 99205, scaling from straightforward to high-complexity medical decision-making. Established patient codes run 99211 through 99215, with 99211 reserved for minimal-service visits that may not require a physician at all. New patient codes carry higher time thresholds than their established-patient counterparts, reflecting the additional work of a first-encounter evaluation.

A coder who knows every code in the range can only bill what the documentation supports. Revenue leakage in E/M coding stems from documentation gaps that leave correctly selected codes vulnerable to downcoding or denial. A practice billing 99213 when the encounter clearly supported 99214 loses a material amount per visit in Medicare reimbursement, multiplied across every provider, every day, without a single claim being denied.

No alert fires. No report surfaces it. Practices invest heavily in coder training, then discover the hidden cost is upstream, at the moment of care, when the encounter note does not capture the clinical complexity that justifies the code.

No amount of coder expertise can recover MDM complexity that was never documented. But before any code in the 99215 range can be applied correctly, there is one foundational question every claim must answer first. That distinction is where we start next.

Key takeaways

• 99202-99205 covers new patients; 99211-99215 covers established patients, those ten codes handle virtually every outpatient office visit a practice bills, and knowing them is the entry requirement, not the finish line.
• Roughly 45% of E&M visits are miscoded before the claim leaves the practice, meaning the revenue problem lives in documentation, not in code lookup.
• New-versus-established misclassification happens before the physician enters the room, when a chart exists in the system, the default assumption is 'established,' and that quiet assumption costs real money across every encounter where it's wrong.
• Since 2021, providers can choose time-based or MDM-based coding for each encounter, that flexibility created two distinct documentation accountability tracks, not one forgiving system with two entry points.
• An E&M code level is defensible only when at least two of the three MDM components, problem complexity, data reviewed, and risk of management, meet or exceed the threshold for that level.
• Undercoding is invisible to payers and never triggers an audit flag, which means billing 99213 on a 99214 encounter compounds quietly, visit after visit, until a distribution review forces the loss into plain sight.
• By the time a coder opens a finalized note, every MDM signal that went uncaptured during the encounter is gone, no coding expertise recovers what the documentation never recorded.
• iScribe Health's AI-powered scribing captures MDM-relevant clinical detail in real time, at the point of care, so the documentation that reaches the coder can actually defend the level the encounter supported.

What Is the Difference Between New Patient and Established Patient CPT Codes

When a patient has a chart in your system, the default assumption is that they are established, and the billing follows automatically from there.

The AMA's Three-Year Rule Exactly When a Patient Counts as New

A new patient is one who has not received any professional services from the same physician, or another physician of the same specialty who belongs to the same group practice, within the past three years. That two-part test, specialty and group membership together, is where the rule is most commonly misapplied, and this is also where beginners in medical billing and coding run into genuine difficulty: instructors and textbooks are often inconsistent or outright incorrect on the specialty-and-group-membership prong of the rule, leaving coders to apply a half-understood standard to real claims.

99202 to 99205 vs 99211 to 99215 Which Code Set Applies

New patient codes carry higher relative value units, reflecting the additional clinical work of a first encounter, including history gathering and baseline assessment that an established visit does not require. Getting this foundational split right is the first discipline of coding consistency, an expertise point that iScribe Health's E&M Coding Intelligence is specifically built to reinforce, surfacing the correct code set at the point of note completion, after the AI drafts the encounter summary, rather than leaving the determination to a downstream billing queue where chart assumptions have already calcified.

• Patient Status

Code Range

Key Characteristic

• New

that same figure–99205

No professional service from same physician/specialty/group in past 3 years; higher RVUs reflecting first-encounter clinical work

• Established

99211-99215

Has received professional service from same physician/specialty/group within past 3 years

The Misclassification Error That Costs Practices Money on the First Encounter

A practice billing 99213 instead of 99202 on a qualifying new patient encounter does not just miss the reimbursement on that single claim; it misclassifies the encounter type and repeats that loss at scale. At note completion, iScribe Health's Automated E&M Coding and Real-Time Denial Alerts surface the coded output and flag classification concerns before a claim goes out the door.

For practices already running a supported EHR, iScribe Health's EHR Integration means this workflow slots in without a parallel system to manage, the ambient documentation experience is seamless, and the coding intelligence is applied at the point where it can still prevent the error rather than remediate it after a denial.

How to Choose Between Time-Based and MDM-Based Code Selection

Pros and cons at a glance

✓ Pros

Choose whichever pathway produces more accurate and defensible code level

Time-based coding often reaches higher level for counseling-heavy encounters

MDM complexity coding typically produces stronger result for brief but dense visits

Flexibility spans nearly every care environment a practice operates in

✗ Cons

Created two distinct documentation accountability tracks, not one forgiving system

Flexibility becomes liability the moment documentation fails to keep pace

Time-based coding only auditable when total time captured contemporaneously

MDM risk categories introduce subjective interpretation varying between providers

Which Pathway Pays More and When the Answer Flips

Time-based coding often reaches a higher level for counseling-heavy encounters, so it is the stronger pathway when a significant portion of the visit involves education or shared decision-making. Conversely, when a visit is brief but clinically dense, MDM complexity coding typically produces the stronger result, rewarding the clinical weight of the encounter rather than its duration.

The AMA E/M Coding Guidelines made this flexibility explicit: providers may select the E/M level using either total time or MDM complexity, whichever method is more advantageous and accurately documented. The AMA's same-figure E/M Office Visit Guidelines extended that same flexibility to inpatient, observation, emergency department, and nursing facility settings with the 2023 updates, so the decision space now spans nearly every care environment a practice operates in.

What Actually Counts as Total Time Under the 2021 Rules

Under the AMA E/M Coding Guidelines, it includes pre-visit chart review, the face-to-face encounter, and post-visit documentation and care coordination activities performed on the date of service, but every minute must be contemporaneously recorded to be auditable.

Failing to capture that time record in real time is a compliance risk. iScribe Health's ambient listening and conversational AI captures the encounter as it unfolds, so the time record is built in real time rather than assembled after the fact, eliminating the reconstruction problem that most audit findings trace back to.

The Audit Exposure Each Pathway Carries

Time-based coding is only auditable when total time is captured contemporaneously; any reconstruction after the encounter converts a legitimate claim into an underdocumented one. MDM-based coding carries its own exposure: the "problems addressed" and "risk" categories introduce subjective interpretation that varies between providers and coders, creating inconsistency across a provider cohort even when everyone understands the rules.

iScribe Health addresses both exposure points in the same workflow, structuring MDM documentation at the point of care so that complexity is recorded with the specificity auditors require, not summarized after the fact.

Time Thresholds and MDM Levels for Established Patient Codes 99211-99215

The Established Patient Code Table as a Revenue Map

Code

MDM Complexity

Time Threshold

99212

Straightforward

10-19 minutes

99213

Low complexity

20-29 minutes

99214

Moderate complexity

30-39 minutes

99215

High complexity

40-54 minutes

Each step up the ladder carries a materially higher physician work RVU, a relationship formalized in the CMS Medicare Physician Fee Schedule Final Rule CY 2018 and governing reimbursement structure ever since.

One of the most persistent struggles in high-volume practices is that coding precision degrades precisely when visit pressure is highest. The documentation process cannot keep pace with the clinical complexity of each encounter. That gap compounds quietly across a high volume of visits before it surfaces in a denial or an audit. iScribe Health's E&M Coding Intelligence is built around that specific failure point: improving coding consistency across providers and improving coding precision to directly impact revenue, at the encounter level, every day.

The 99213-vs-99214 Fault Line Where Undercoding Hides

The visit supports 99214. But the note written at 7 p.m. after a full schedule captures only a compressed summary of what occurred.

Through Ambient Listening and Conversational AI, the system captures clinical reasoning in real time, during the encounter, so the note reflects the actual MDM complexity the provider exercised rather than a compressed summary written under end-of-day fatigue.

What Defensible 99215 Documentation Actually Requires

High complexity MDM requires at least one of the following: a decision to hospitalize, a decision involving a drug requiring intensive monitoring, or management of a condition posing a threat to life or bodily function.

Key takeaway: Documentation deficiencies at the point of care are the primary driver of E/M coding inaccuracy.

iScribe Health's Automated E&M Coding, working in concert with Ambient Listening, closes that gap at the source: the system drafts a structured encounter summary that surfaces the specific elements, drug name, risk category, clinical rationale, that make a that same figure defensible on review.

Time Thresholds and MDM Levels for New Patient Codes 99202-99205

Selecting the right code from 99202 through 99205 hinges on two independent pathways: medical decision-making complexity and total time spent on the date of the encounter, and confusing the two or miscounting either one is where new patient visits most often land at the wrong level. Total time extends beyond the face-to-face visit to include pre-visit chart review and post-visit care coordination, so providers who document only what happens in the room routinely leave a higher-level code unearned. Each threshold covered below maps the specific MDM level and minute range that qualifies, along with where contemporaneous documentation, including the kind captured through ambient AI, makes the difference between a defensible code and a missed one.

Four new patient office visit codes carry distinct time and complexity requirements, and confusing them costs real money on every miscoded claim.

1. 99202 - Straightforward MDM, 15-29 Minutes Total Time

According to the AMA CPT E/M Guidelines, that same figure requires either straightforward medical decision-making or 15-29 minutes of total time on the date of the encounter. Total time includes pre-visit chart review and post-visit care coordination, not just face-to-face minutes. This is the floor for new patients after CPT that same figure was retired in 2021. Practices that count only face-to-face minutes systematically undercount billable time at that same figure, leaving a legitimate 99203 on the table. Because new patient encounters in complex or specialist settings require more pre-visit preparation, the combined time across all encounter phases frequently crosses higher thresholds, but only when documented contemporaneously rather than estimated retrospectively.

This is precisely where ambient AI documentation changes the equation. iScribe Health's Ambient Listening and Conversational AI captures the full encounter in real time, including the clinical reasoning that precedes and follows the face-to-face visit, so that the time entered into the note reflects what the provider spent rather than a compressed estimate added at the end of a twelve-hour shift.

2. 99203 - Low Complexity MDM, 30-44 Minutes Total Time

99203 applies when the encounter reaches low complexity MDM or 30-44 minutes of total time. A new patient presenting with one or two stable chronic conditions, requiring medication review and updated baseline assessment, typically lands here. The documentation burden is heavier than it looks: low complexity MDM still requires capturing the number and nature of problems addressed, data reviewed, and risk level. Practices that benchmark this against established patient norms systematically underbill, because first-encounter history-gathering adds genuine clinical minutes that must be recorded contemporaneously to count.

Physicians in high-volume practices, those regularly charting significant hours outside of patient care time, are most exposed to this pattern. When post-visit documentation is completed from memory at the end of the day, the granular detail that justifies 99203 over 99202 erodes. iScribe Health's AI-powered scribing drafts the encounter summary at the point of note completion, after the AI has listened to the full encounter, so low-complexity MDM components are captured in the moment they occur rather than reconstructed later.

3. 99204 - Moderate Complexity MDM, 45-59 Minutes Total Time

Moderate complexity MDM or 45-59 total minutes defines 99204, the code where under-billing risk concentrates most sharply for new patient visits. A patient presenting with multiple chronic conditions requiring prescription drug management sits squarely here. The gap between 99204 and its established-patient counterpart is material at the reimbursement level, making accurate documentation of MDM components the highest-leverage action a practice can take per encounter.

iScribe Health's E&M Coding Intelligence applies at exactly this threshold. Because the AI drafts the note from ambient listening rather than a provider's end-of-day recall, the MDM data elements, number and nature of problems, data reviewed, risk of complications, appear in the note as they were clinically generated. Real-Time Denial Alerts then flag coding inconsistencies before the claim leaves the practice, reducing the downstream cost of a miscoded that same figure. For high-volume practices and health systems, that per-encounter accuracy compounds across every day of clinical practice.

4. 99205 - High Complexity MDM, 60-74 Minutes Total Time

99205 requires high complexity MDM or 60-74 total minutes, covering new patients whose care involves severe exacerbations, drug therapy requiring intensive monitoring, or decisions with significant morbidity risk. A complex chronic disease consult reaching the high-complexity time threshold with high-risk MDM documentation supports 99205, but only if total time across all encounter phases is recorded in the note. This is also where 99223 confusion surfaces: coders occasionally misapply that initial hospital admission code to a complex office visit. CPT 99223 is a hospital care code, and billing it for an outpatient encounter is a structural error, not a complexity judgment call.

Documenting the full arc of a that same figure encounter, pre-visit chart review, the visit itself, and post-visit care coordination, is the kind of multi-phase time capture that manual workflows compress. iScribe Health's Automated E&M Coding and EHR Integration mean the code suggested at note completion is grounded in documented time and MDM elements rather than a provider's best guess under end-of-day fatigue, and it flows directly into the supported EHR without a separate data-entry step.

New patient encounters demand the most documentation per visit, yet practices often rely on the same manual workflow they use for returning patients. The clinical minutes spent on pre-visit chart review and post-visit coordination are real and billable, as multiple industry sources underscore when examining how unrecorded physician time systematically distorts encounter-level reimbursement, but they count only when captured in real time. iScribe Health's ambient AI documentation reduces the time physicians spend on post-visit documentation so they can see more patients per day without burning out, and it ensures the time recorded in the note reflects what the provider spent, so the code billed can be defended at the level the encounter genuinely supports.

Knowing which code number applies to a new patient visit is the necessary starting point, but the code is only as defensible as the documentation behind it. The next section breaks down exactly what coders must capture across all three MDM components to make that documentation hold up when a payer looks twice.

What the Three MDM Components Actually Require Coders to Document

Three elements determine whether a given E/M code level is defensible: the number and complexity of problems addressed, the amount and complexity of data reviewed and ordered, and the risk of complications or morbidity. Under the AMA's E/M guidelines, a code level is supported only when at least two of the three components meet or exceed the threshold for that level. Knowing that framework exists is not the same as documenting the clinical reasoning that satisfies it.

1. Number and Complexity of Problems Addressed - The Foundation of Every Office Visit CPT Code Selection

The AMA's definition draws a hard line between a condition actively managed during the encounter and one merely listed in the note. The problems-addressed component is the most frequently cited deficiency in E/M audit findings precisely because clinicians document the diagnosis but compress the management reasoning into shorthand that payer reviewers cannot credit.

2. Amount and Complexity of Data Reviewed and Ordered - The Most Objectively Scoreable MDM Pillar

A provider who reads an outside cardiology report must document that interpretation explicitly, not simply note that the record was present. The gap between "reviewed" and "interpreted" is where many and 99215 claims lose their footing during payer audits, because the scoring threshold is concrete and the documentation shortfall is measurable.

3. Risk of Complications and/or Morbidity or Mortality - The Pillar Most Likely to Fail Audit When Documentation Is Thin

Adjusting a high-risk medication or deferring a specialist referral qualifies as a moderate-to-high risk decision, but only when the note records the clinical reasoning behind each choice.

Related Reading

• Cpt Code 99244
• Cpt Code 99203
• Cpt Code 99232
• Cpt Code 99221
• Cpt Code 99205
• Cpt Code 99202
• Cpt Code 99223
• Cpt Code 99222
• Cpt Code 99233

Why Getting the Code Number Right Still Leaves Most Practices Exposed to Revenue Leakage

Most practice administrators and billing decision-makers believe that once coders know the correct CPT code ranges and time thresholds, E/M coding accuracy is essentially solved and revenue leakage is a residual, manageable exception. That assumption, however, consistently fails under scrutiny.

The 45% Miscoding Rate That Persists Even in Educated Practices

Study after study on E/M coding accuracy points to the same uncomfortable finding: a significant share of office visit claims are miscoded before they ever reach the payer. That miscoding is not random noise distributed evenly across levels. Undercoding is the dominant error pattern in primary care and specialty practices alike, and it runs systematically through provider cohorts rather than appearing as isolated one-off mistakes. The problem is documentation fidelity.

Key takeaway: Physician work embedded in E/M codes carries RVU-based value comparable to procedure and test codes in the CMS system, so undercoding one E/M level across a cohort of providers is a direct, dollar-for-dollar revenue loss proportional to visit volume.

Provider Confidence vs. Actual Accuracy in Orthopedic Coding

In a Journal of Healthcare Finance iScribe audit study of 941 orthopedic encounters, iScribe's analysis found a 33% overcoding rate and only 54.8% provider-AI agreement on code selection. That confidence-accuracy gap is where silent revenue leakage originates.

Why Errors Compound Silently - The Reactive Discovery Problem in E/M Billing

The structural danger is the velocity at which systematic errors accumulate before they surface. If a single provider sees 20 established patients per day and systematically bills that same figure instead of a defensible 99214, the per-visit RVU difference, confirmed by peer-reviewed research to carry real dollar-for-dollar RVU weight in the CMS system, multiplies across thousands of encounters before anyone flags it.

How AI-Assisted Documentation Closes the E/M Coding Accuracy Gap at the Encounter Level

Every MDM signal that went uncaptured during that visit, the complexity of problems addressed, the data reviewed, the risk of patient management discussed, is gone. No amount of coding expertise recovers what the documentation never recorded.

Why Post-Submission Coding Review Cannot Recover Undocumented MDM Complexity

As broader industry trends make plain, documentation deficiencies, not coder knowledge gaps, are the root cause of E/M coding inaccuracy. When the note fails to reflect the full clinical picture, even a skilled coder cannot assign a higher-complexity code the documentation does not support. MDM-based code selection requires meeting thresholds in at least two of three elements, but the element most frequently underdocumented, complexity of problems addressed and risk of patient management, is also the element that most directly reflects what actually happened in the encounter.

Practices that treat MDM documentation as a post-encounter narrative task are systematically converting clinically legitimate higher-level visits into defensible lower-level claims, with no audit or coder review process capable of recovering the lost complexity. Training coders on MDM thresholds is a downstream intervention, and by the time it is applied, the opportunity to document what actually occurred has already passed.

What Contemporaneous Documentation Must Capture to Satisfy MDM Thresholds

The only point at which MDM complexity can be accurately recorded is during the encounter itself. Rather than adding a billing layer after the fact, ambient real-time clinical documentation captures MDM complexity signals as care is delivered, problems addressed, data reviewed, risk discussed, so that when the provider signs the note, the complexity actually delivered is already present and defensible.

What most teams report is that real-time documentation tools are the mechanism that converts coding accuracy from a downstream correction effort into a natural output of the visit itself. When a tool flags documentation gaps before the note is signed, the practice gains something no post-submission audit can offer: the ability to act while correction is still possible. A practical pre-signature checklist reflects that principle directly:

• The problems-addressed element is documented with enough specificity to defensibly support the target code level
• Data reviewed during the encounter is captured in the note, not reconstructed after the fact
• Risk of patient management, including drug names, risk categories, and clinical rationale, is recorded before the note is locked
• Documentation gaps are flagged to the provider before the note is signed, while correction is still possible

The documentation-accuracy problem is solvable, but only if the solution is applied at the encounter level, before the note is locked and the coding window closes.

Related Reading

• Cpt Code 99499
• Cpt Code 99386
• Cpt Code 99395
• Cpt Code 99385
• Cpt Code 99284
• Cpt Code 99396
• Cpt Code 99283
• Cpt Code 99024

Next steps

If your practice is billing 99213 across encounters that clinically support 99214, the path forward starts with recognizing that the pattern is invisible until a distribution review or audit forces it into view. Start with our AI medical scribe.

The 99213-versus-99214 boundary is the single highest-stakes coding decision by volume in most practices, and documentation drag at the point of care is what keeps clinically defensible visits coded at the lower level. That means post-submission coder review cannot fix it, because coders can only support what the note already says. The 2021 dual-pathway rule compounds the exposure: time-based and MDM-based selection are only advantageous when the corresponding documentation was captured in real time, not reconstructed from memory at the end of a packed schedule. Together, those two realities point to a single intervention: shifting documentation capture to the moment of care, before the note is finalized and the MDM complexity window closes.

Start with AI medical scribe technology that records problems addressed, data reviewed, and risk considerations while the visit is still happening. After iScribe's ambient AI drafts the encounter summary, E&M Coding Intelligence surfaces the appropriate code level based on what was actually documented in the room, and Real-Time Denial Alerts flag gaps before a claim leaves the practice. For practices already running a supported EHR, that workflow slots in without a parallel system to manage, so the accuracy improvement is realized across every patient encounter and every day of clinical practice.

Frequently Asked Questions

Is CPT code 99201 still valid for billing new patient visits?

No. CPT 99201 was retired effective January 1, 2021, when the AMA overhauled the E/M guidelines. Any claim submitted with 99201 will be rejected, making 99202 the lowest billable code for new patients today.

When does a patient with an existing chart actually count as a new patient?

A patient counts as new if they have not received any professional service from the same physician, or another physician of the same specialty in the same group practice, within the past three years. Both specialty and group membership matter, a patient seen by a cardiologist in the same group three years and one day ago qualifies as a new patient for any other specialty in that group, and billing them as established on that encounter is a misclassification.

What's the difference between billing 99213 and 99214 for an established patient?

The key difference is the level of clinical complexity, or total time, the documentation supports. A practice billing 99213 when the encounter clearly supported 99214 loses a material amount per visit in Medicare reimbursement, multiplied across every provider and every day, without a single claim being denied and without any alert firing to surface the pattern.

Does total time for time-based E/M coding include only the face-to-face visit?

No. Under the AMA E/M Coding Guidelines, total time includes pre-visit chart review, the face-to-face encounter, and post-visit documentation and care coordination activities performed on the date of service, all three components must be captured together to make a time-based claim defensible. If the note only reflects in-room time, the documentation does not support a code that relies on the combined total.

Which coding pathway, time-based or MDM, should I use to get the higher code level?

You should choose whichever pathway your documentation already supports, not the one that yields the higher code in the abstract. For counseling-heavy encounters, time-based coding often reaches a higher level; for visits that are brief but clinically dense, MDM complexity typically produces the stronger result. Treating the selection as a post-encounter decision, after the note is finalized, converts a legitimate revenue opportunity into an underdocumented claim.