CPT Code 99203 Rules and Reimbursement Guide 2026
CPT code 99203 rules for 2026, billing teams. Automate E&M coding, close the visibility gap, and recover reimbursement across every new-patient encounter.

Most 99203 miscoding is not a training problem. It is a visibility problem, and the revenue it quietly erases compounds across every new-patient encounter your billing team thinks it already has right.
Most practice administrators and billing decision-makers believe their 99203 billings are accurate as long as their coders know the code's MDM and time thresholds, treating miscoding as a knowledge gap that training or a better fee schedule can fix. That assumption deserves scrutiny. Only about half of all E&M visits are correctly documented on the first attempt.
That figure, drawn from research published in the Muma Business Review in February 2022, should stop any practice administrator cold. For a code that sits at the center of nearly every new-patient encounter workflow, CPT 99203 is far less routine than most billing teams treat it. Understanding precisely what it covers, and where its boundaries sit, is the first line of defense against revenue loss that compounds quietly across hundreds of encounters.
See our AI medical scribe for how this works in practice.
"Students learning medical billing and coding struggle to understand when and why CPT code 99203 (a new patient, level 3 E/M service) would require Modifier 25, indicating a gap in practical education around this specific code."
— what we hear from medical billing and coding students
CPT code 99203 is a Level 3 office or other outpatient visit for a new patient. Per the American Medical Association's E/M coding guidelines, the code requires either low-complexity medical decision making (MDM) or a total time of 30 to 44 minutes on the date of the encounter. That "or" carries real weight: a provider meets the threshold through one pathway, not both simultaneously.

The MDM pathway demands that the clinical note demonstrate low-complexity problem complexity, appropriate data review, and a low risk of complications. The time pathway requires the note to explicitly state total time, including pre-visit preparation and care coordination performed on the same date, not just face-to-face minutes. A note that asserts low complexity without the clinical narrative behind it is not defensible. That distinction is where most practices quietly lose ground. The AMA's own guidance identifies two prototypical presentations for this code: a new patient with a single stable chronic illness, such as well-controlled hypertension or type 2 diabetes, and a new patient with an acute uncomplicated injury or illness, such as a minor laceration or an upper respiratory infection.
What pushes a visit outside that same figure's boundaries is equally instructive. A new patient presenting with two or more chronic conditions requiring active management, or a problem carrying moderate risk, shifts the encounter toward 99204. Likewise, if the provider's total time reaches 45 minutes, the correct code is 99204, not 99203. Defaulting to that same figure when the clinical picture has crossed that line is underbilling, and it happens routinely when documentation doesn't fully capture what occurred.
CPT that same figure is the most commonly billed mid-level new-patient E&M code, making its accurate assignment critical to both compliance and a practice's financial performance, according to industry data. A pattern of miscoded encounters across a multi-provider group, compounding week over week, becomes a measurable annual revenue gap or an audit liability that most practices never detect until a payer flags it. The gap is not effort.
The gap is visibility. AI medical scribe technology addresses this by capturing MDM-relevant clinical detail at the point of care, before the note is closed, so the code rests on a complete narrative rather than a reconstructed one. But before a practice can bill that same figure correctly, it has to answer a deceptively simple question that trips up multi-provider groups every day: who, exactly, counts as a new patient?
Key takeaways
- Only about half of all E&M visits are correctly documented on the first attempt, meaning a practice billing 99203 confidently is, statistically, miscoding roughly half the time.
- The most expensive assumption in outpatient E&M coding is that a patient who has never seen this doctor is automatically a new patient, the definition is more specific than that, and getting it wrong cuts both ways.
- CPT 99203 sits between two real billing errors: undercoding a visit that qualifies for 99204, and overcoding one that only supports 99202, neither is a conservative choice, both carry audit exposure.
- A single 99203 reimbursement can range from roughly $60 to more than $126 depending on payer, geography, and facility setting, that gap compounds across every new-patient slot on the schedule.
- Auditors do not look for a code label; they look for a clinical narrative that earns it, a note that asserts low-complexity MDM without the documentation to back it up fails the same way a blank note does.
- The documentation gap does not form at the billing desk, it forms in the exam room, before any coder opens the chart, which is why coder training alone cannot close it.
- iScribe Health's Automated E&M Coding generates E&M codes directly from clinical documentation in real time, capturing what happened in the encounter before the note is finalized, closing the gap at the point where it actually opens.
New Patient Definition for CPT 99203 and Why It Gets Misapplied
That visibility problem compounds the moment a practice turns to the new-patient question, because the most expensive assumption in outpatient E&M coding is also the most intuitive one: if a patient has never seen this doctor, they must be new. Practices applying that logic without verifying whether the patient actually meets the new-patient definition are not cutting corners; they simply lack the cross-provider visibility to know the question needs asking at all.

The AMA Three-Year Rule - What New Patient Actually Requires Under CPT Guidelines
A new patient, under American Medical Association (AMA) CPT Guidelines, is one who has not received any professional services from the physician, or any physician of the same specialty and subspecialty who belongs to the same group practice, within the past three years. Three conditions must all be true before a practice can bill a new-patient code: the patient has no prior encounter within the three-year window, no provider in the same specialty within the group saw them during that period, and the visit itself meets the documentation thresholds for a new-patient level-three service. A patient returning to a different orthopedist in the same group after two and a half years is not a new patient under CPT guidelines, regardless of whether the treating physician has ever met them.
How Group Practice Structure Silently Disqualifies New Patients
The structural risk lives at the Tax Identification Number level, where a patient seen by Physician B in the same cardiology group 26 months ago is already established, even if the billing provider has no awareness of that prior encounter. iScribe Health's E&M Coding Intelligence and Real-Time Denial Alerts are designed to close exactly this gap.
The Audit Exposure When 99203 Is Billed for Established-Patient Encounters
New-patient codes reimburse meaningfully higher than established-patient codes, which makes the delta both the incentive that drives the error and the dollar amount that must be repaid when an audit surfaces it. iScribe Health's Automated E&M Coding and Real-Time Denial Alerts work together to reduce this exposure at scale. The compliance risk is structural, and it scales with the size of the group. iScribe Health's ambient AI documentation layer is built to scale with it.
Medical Decision Making and Time Requirements for 99203
This section breaks down both qualifiers, including the specific MDM elements that trip up even experienced providers, so the clinical work your physicians perform translates into claims that hold up.

Two Thresholds, Two Failure Modes
Missing either threshold independently is enough to disqualify the code.
Low-Complexity MDM as the Clinical Floor for 99203
Qualifying problem types include one self-limited or minor problem, or one stable chronic illness with prescribed therapy.
This is precisely where high patient volumes make the problem worse. iScribe's ambient AI medical scribe is most beneficial when clinicians are seeing high patient volumes and spending significant time on after-hours documentation, exactly the conditions under which MDM elements go unrecorded not because the provider doesn't understand them, but because there is no time to articulate them during or immediately after the encounter. iScribe Health's E&M Coding Intelligence and Automated E&M Coding capabilities evaluate documentation at the point of note completion, after the AI drafts the encounter summary, so the MDM elements are surfaced and confirmed before the chart is locked rather than discovered missing during a post-payment audit.
30 to 44 Minutes of Total Time on the Date of Service
Critically, total time is not limited to face-to-face encounter time.
For clinicians already charting two or more hours outside of patient care time, this gap is a daily revenue leak. iScribe Health's Ambient Listening / Conversational AI captures pre-visit and post-visit activity that qualifies under the time pathway, reducing the after-hours documentation burden while simultaneously building a more complete time record. The result is a time record that is both accurate and audit-ready without requiring additional effort from the provider.
MDM vs. Time as a Revenue Strategy Decision
Choosing between the MDM and time pathways carries direct revenue consequences. When a visit's clinical complexity genuinely supports low-complexity MDM, defaulting to time-based billing without explicit minute documentation creates audit exposure that MDM-based billing would have avoided. The inverse is equally costly: a provider who spends 35 minutes with a new patient but documents only "low MDM" without noting time loses the time pathway entirely, and the visit risks being downcoded.
iScribe Health's coding intelligence evaluates both pathways simultaneously and surfaces the stronger supported option before the note is finalized. Because iScribe Health integrates directly with supported EHRs through its EHR Integration capability, this intelligence surfaces inside the clinical workflow physicians already use, with no separate tool to open and no extra step to remember.
Understanding how these thresholds interact across adjacent codes is the next step.
How 99203 Compares to 99202 and 99204 for New Patient Visits

99202 vs. that same figure - Where the Floor Is and When You've Already Crossed It
In high-volume practices where clinicians regularly chart two or more hours outside of patient care time, encounter notes are frequently reconstructed from memory well after the visit ends, and reconstructed notes compress clinical detail, which is exactly how low-complexity MDM gets collapsed into a straightforward code. iScribe Health's Ambient AI Documentation captures clinical notes during or after the patient encounter using AI scribing, so the MDM elements that justify that same figure, the second self-limited condition, the lab order, the external record review, are present in the note at the point of note completion, not selectively recovered later.
When to Bill 99204 Instead of 99203 - Moderate MDM or 45 Minutes
A new-patient visit running 46 minutes, documented only as that same figure, represents a meaningful per-visit revenue loss at Medicare rates.
iScribe Health's Automated E&M Coding evaluates clinical documentation at the point of note completion, after the AI drafts the encounter summary, flagging in real time when documented MDM complexity or total time crosses the 99204 threshold, so the revenue that belongs to a higher code level is not quietly abandoned visit after visit.
The Center for Sports Medicine & Orthopaedics, Chattanooga audit, where 941 encounters were analyzed, a 54.8% provider-AI agreement rate was found, and a 33% overcoding rate was uncovered under a traditional post-visit process, illustrates why the point of capture, not the billing desk, is where coding accuracy is determined.
Because iScribe Health integrates directly with supported EHRs, the ambient documentation and E&M coding intelligence materialize inside the workflow the practice is already running, with no parallel charting environment and no double-entry.
99203 vs. 99213 - Why These Two Codes Are Never Interchangeable
99203 applies exclusively to new patients. CPT 99213, by contrast, is an established-patient code with a 20 to 29-minute total time threshold, per the AMA 2021 E/M Guidelines.
A patient who "hasn't been seen in years" may still qualify as established under the AMA's three-year rule, as defined in the AMA CPT Code Set (2021 E/M Revisions). The codes share no interchangeable logic: different patient status, different time windows, different MDM anchors. iScribe Health's Ambient Listening and Conversational AI captures the contextual detail of each encounter, including patient history signals relevant to new versus established status, so the note supports the correct patient classification from the moment documentation is complete, rather than leaving that determination to a downstream coder working from an incomplete summary.
Related Reading
- Cpt Code 99233
- Cpt Code 99232
- Cpt Code 99244
- Cpt Code 99221
- Cpt Code 99205
- Cpt Code 99223
- Cpt Code 99202
- Cpt Code 99222
2026 Reimbursement Rates for CPT Code 99203 Across Payers
The actual payment for a single CPT that same figure encounter ranges from roughly $60 to more than $126 depending on payer type, geographic location, and facility setting. That spread is a calculable revenue gap that compounds across every new-patient visit on the schedule.
Here is what the numbers look like across the three payer categories that matter most to a practice's monthly revenue report.
1. Medicare National Non-Facility Rate for CPT Code 99203: ~$117.57 in 2026
Medicare's $117.57 national non-facility rate is the most cited benchmark for that same figure reimbursement, but it is a starting point. Per the CMS-1832-F final rule published November 2025, actual payment is calculated by multiplying the code's relative value units by the conversion factor and then adjusting for Geographic Practice Cost Indices (GPCIs). That GPCI adjustment means a practice in a high-cost metro can collect meaningfully above $117.57, while a rural practice may land below it for the same encounter.
Facility setting compounds this further: billing that same figure in a hospital outpatient department rather than an office setting triggers a lower facility rate, sometimes by $20 or more per visit. The American Society of Nuclear Cardiology (ASNC) analysis of the 2026 proposed rule also flags that CMS is proposing different conversion factors for different clinician types, meaning the billing provider's classification, not just the code, shapes the final dollar.
One of the most consistent problems practices face at this layer is coding inconsistency across providers. When two clinicians see clinically similar new patients and one routinely documents to a that same figure standard while the other undershoots to 99202, the revenue difference is invisible in the encounter count but fully visible in the monthly collections report. iScribe Health's E&M Coding Intelligence addresses this directly: at the point of note completion, after the AI drafts the encounter summary, it surfaces an automated E&M code suggestion drawn from the clinical documentation as it was captured, not reconstructed from memory after a rushed finalization. That process improvement works across every provider on the schedule, which is where coding consistency either compounds revenue or quietly erodes it.
2. Medicaid Median Reimbursement for CPT Code 99203: ~$90.52 Nationally
Medicaid reimbursement for CPT that same figure varies considerably across state programs, with published state fee schedules showing a wide band of per-encounter rates. For practices with high Medicaid volume, that spread is a direct determinant of whether the payer mix is sustainable.
Practices in low-paying states often discover this gap only after a revenue cycle review, not during contract negotiations, which is where the damage is hardest to reverse. The reimbursement integrity problem at the Medicaid tier is compounded by documentation: if the clinical note does not defensibly support that same figure, if the medical decision-making or history elements fall short under payer review, the practice is at risk of collecting nothing, or of absorbing a denial and the administrative cost of an appeal.
iScribe Health's Ambient AI Documentation captures the encounter in real time through Ambient Listening and Conversational AI, so the note reflects what actually happened in the room rather than what the clinician could recall after completing a full schedule. That distinction separates a note that holds up from one that does not.
3. Commercial Payer Average for CPT Code 99203: $116-$126 per Encounter
Commercial payers cluster above Medicare for that same figure, but the range across major carriers is wider than most administrators expect. Commercial rate benchmarks place major national carriers, BCBS, Anthem, United, Cigna, and Aetna, in a cluster above Medicare for that same figure, with per-encounter rates generally ranging from the mid-$110s to the mid-$120s depending on contract vintage and geographic market. Smaller or newer practices negotiating contracts without benchmark data often accept rates at or below the Medicare floor, leaving $5 to $10 per encounter uncollected. At 200 new-patient visits per month, that gap reaches $12,000 to $24,000 annually before accounting for any undercoding at the encounter level.
Real-Time Denial Alerts, surfaced within iScribe Health's workflow at the point of note completion, give practices an additional layer of reimbursement integrity at the commercial tier. Rather than discovering a documentation gap during a payer audit weeks after the encounter, the practice receives the alert while the encounter is still fresh and correctable. This is particularly impactful in high-volume practices or health systems where clinicians regularly chart two or more hours outside of patient care time; ambient documentation shortens that tail, and denial alerts prevent the revenue loss that typically hides inside it.
Because iScribe Health integrates directly with supported EHRs, the coding intelligence and denial alerts surface inside the existing workflow without requiring a separate platform or manual data transfer, so reimbursement and payment workflows are simplified across every encounter.
Most practice administrators track fee schedule updates and assume that awareness of the published rate is enough to protect revenue. Reimbursement optimization for that same figure requires more than fee schedule awareness, because the actual dollar value of the code varies by clinician type, geographic location, and facility setting, and that variable payment is only realized if the documentation defensibly supports the level billed. A practice monitoring conversion factor changes while leaving its documentation capture process unreliable is optimizing the wrong variable.
The ASNC 2026 fee schedule analysis makes the ceiling problem plain; the published rate only materializes if the clinical note holds up under payer review. That is where iScribe Health's automated scribing closes the gap: by generating E&M code suggestions from the clinical documentation as it is captured, not after a rushed finalized note, it ensures that every defensible that same figure encounter reaches the rate the payer schedule already authorizes.
These rate differences make one thing clear: the published fee schedule is the ceiling, and a practice reaches it only when every defensible encounter is documented completely enough to withstand payer review.
99203 Documentation Requirements and 2024-2026 Rule Updates

What the Note Must Actually Contain to Earn a 99203
According to the AMA CPT E/M Office Visit Guidelines (2021 Revision), a note that simply asserts "low complexity MDM" without a supporting clinical narrative is not defensible under audit because reviewers evaluate whether the documented content actually meets the criteria, not whether the provider labeled it correctly.
This is precisely the kind of structural capture failure that the Journal of Medicine, Health Sciences Research (JMHSR), 'Evaluating the Quality and Completeness of Inpatient Medical Record Documentation: A Clinical Audit' identifies as pervasive across clinical settings: documentation gaps are recurring patterns tied to how and when notes are composed. In high-volume settings where clinicians regularly chart two or more hours outside of patient care time, the problem complexity, data sources, and risk language that existed in the room simply never make it to the record.
iScribe Health's ambient AI documentation addresses this at the source, capturing the clinical narrative in real time during the encounter, so the content needed to support low-complexity MDM is present before the note is finalized, not reconstructed afterward. At the point of note completion, after the AI drafts the encounter summary, iScribe Health's E/M Coding Intelligence evaluates whether the documented problem complexity, data reviewed, and risk language actually meet the threshold, flagging any element that falls short before the claim is submitted. This closed-loop review confirms defensibility at the moment of documentation, before a post-submission audit surfaces the gap.
How to Document by Time Without Leaving the Record Exposed
When using the time pathway, the note must explicitly state the total minutes spent on the date of service, inclusive of non-face-to-face work such as reviewing records, ordering tests, and completing documentation. A provider who spent 38 minutes but documented nothing beyond the clinical encounter has no time-based claim to stand on.
What 2024 to 2026 Rule Updates Actually Changed
Under the CMS Calendar Year 2024 Physician Fee Schedule Final Rule, when a new-patient visit is shared between a physician and a non-physician practitioner, the note must include an attestation identifying who performed the substantive portion of the visit. Absent that attestation, the claim is exposed regardless of how well the clinical content is otherwise documented.
The OIG Work Plan and CMS Evaluation and Management Audit Guidelines confirm that documentation non-compliance is a persistent systemic problem, and enforcement activity around these requirements has continued to intensify through the 2025-2026 review cycle.
iScribe Health's Real-Time Denial Alerts surface missing attestation requirements and documentation gaps before submission. For practices already operating on a supported EHR, iScribe Health's EHR Integration means this happens within the existing workflow: the ambient documentation experience captures the encounter, the AI drafts the note, E/M Coding Intelligence evaluates defensibility, and denial alerts flag any outstanding requirements, all without pulling the physician out of patient care.
- 1 — Problem Complexity — Explicit identification of problem type (e.g. single stable chronic illness; one acute uncomplicated illness/injury) with supporting clinical detail, not a label alone
- 2 — Amount and Complexity of Data Reviewed — Named data sources reviewed or ordered (e.g. prior records, labs, external notes) with a brief statement of relevance
- 3 — Risk of Complications or Morbidity — Documented risk level with clinical rationale (e.g. prescription drug management, referral decision, or risk stratification language), not a generic phrase
- 4 — Time (if time pathway used) — Explicit total minutes on date of service (30-44 min for that same figure), inclusive of non-face-to-face work; "lengthy visit" is insufficient
- 5 — Split/Shared Attestation (if applicable) — Attestation naming the provider who performed the substantive portion, per the CY 2024 Physician Fee Schedule Final Rule
The Hidden 99203 Audit Risk Most Practices Discover Too Late
Federal audit data makes the stakes concrete: an OIG audit found that one organization accumulated $3.3 million in overpayments across just two years, with the root cause identified not as coder ignorance but as claims billed without adequate supporting documentation. That distinction matters, because it means the risk is not confined to poorly run practices.

One Week, One Orthopedic Group and a 33% Overcoding Rate Nobody Saw Coming
The OIG's broader audit record confirms this pattern: practices are not self-identifying miscoding errors before external review, which means liability accumulates silently until a payer or federal auditor surfaces it. By that point, repayment demands, interest, and potential penalties are already in motion.
The Invisible Gap When the Visit Justifies the Code but the Note Doesn't
The core problem is that a code is only as defensible as the documentation behind it. If the clinical narrative omits problem complexity, skips data reviewed, or reduces the risk assessment to a generic phrase, the code becomes indefensible regardless of what actually happened in the room.
Related Reading
- Cpt Code 99284
- Cpt Code 99396
- Cpt Code 99499
- Cpt Code 99386
- Cpt Code 99395
- Cpt Code 99385
- Cpt Code 99283
- Cpt Code 99024
How Automated E&M Coding Closes the 99203 Documentation Gap
Every coder who has flagged a misleveled that same figure after the note was finalized has already lost the fight, because the documentation that determined the code was already complete and closed. The gap did not form at the billing desk; it formed in the exam room.

Why the Gap Forms at the Point of Care Not at the Billing Desk
Physicians document under pressure, and the note is often completed from memory after the encounter has ended.
Recall-based documentation compresses clinical detail. Problem complexity gets summarized. Data reviewed goes unmentioned.
Risk of complications never makes it into the record, and the resulting code reflects what was written rather than what actually occurred during the visit.
Post-visit audits find errors. They cannot prevent them.
How Ambient AI Surfaces MDM Evidence Before the Note Closes
Real-time ambient capture changes the sequence entirely. What most teams report bears out structurally: ambient AI captures MDM-relevant detail, including problem complexity, data reviewed, and risk, in real time before the note is closed, ensuring the resulting code has full clinical narrative behind it rather than reconstructed or abbreviated post-visit documentation.
How Real-Time Documentation Addresses Both Undercoding Revenue Loss and Overcoding Audit Risk
Most interventions fix one side of the miscoding problem. Training reduces overcoding, while retrospective auditing recovers some undercoded revenue after the fact. Neither addresses both simultaneously. Across the market, the structural argument holds directly: closing the documentation capture gap simultaneously eliminates undercoding revenue loss and overcoding audit exposure, making real-time ambient documentation a single solution to two distinct financial risks that training or retrospective auditing each address only partially.
Automated E&M coding through iScribe Health's coding intelligence surfaces the defensible code level before the chart is locked, giving practice administrators both the revenue recovery and the audit protection that post-visit review can only partially approximate.
The next section walks through a targeted mini-audit that surfaces how much that same figure revenue your current documentation workflow is leaving on the table, and what a corrected baseline looks like before you change a single workflow.
iScribe Health is built for independent specialty practices and mid-size groups where documentation capture happens at the point of care without a dedicated in-house scribe program. It is not designed for large hospital systems or academic medical centers that already operate enterprise-level EHR build teams and embedded scribe workflows at scale, where the implementation model and integration requirements would be a poor fit.
Next steps
If your 99203 encounters are slipping between defensible and indefensible based on what made it into the note rather than what happened in the room, the path forward starts with fixing the capture process before the chart closes, not reviewing claims after they leave. Start with our AI medical scribe.
Knowing the MDM thresholds for 99203 is necessary but structurally insufficient, because a provider who understands the code perfectly will still generate a non-defensible claim whenever problem complexity, data reviewed, and risk of complications are absent from the clinical narrative. That is the knowledge gap and the capture gap operating as two distinct failure modes at once. At the same time, conservative coding drift from that same figure to 99202 is self-concealing, because the same risk aversion that causes undercoding also suppresses the internal measurement that would reveal it, masking cumulative annual revenue loss while the visit content would have defensibly supported a higher level.
Together, they point to one corrective action: capturing clinical detail in real time, at the point of care, before memory compresses what the record needs to hold.
Start with an AI medical scribe built for the clinical workflow your practice already runs. After the AI drafts the encounter summary, E&M Coding Intelligence evaluates whether the documented MDM elements meet the 99203 threshold or defensibly support 99204, and Real-Time Denial Alerts flag gaps before the claim is ever submitted. The documentation burden drops, the note reflects what actually occurred, and every defensible encounter reaches the reimbursement the payer schedule already authorizes.
Frequently Asked Questions
How long does a 99203 visit need to be?
When using the time pathway, CPT 99203 requires 30 to 44 minutes of total time on the date of service. Total time includes pre-visit preparation, reviewing prior records, ordering tests, documenting the care plan, and care coordination performed on the same calendar date, not just face-to-face minutes.
What level of medical decision making is required for 99203?
CPT 99203 requires low-complexity medical decision making as its minimum MDM level. All three MDM elements, problem complexity, data reviewed, and management risk, must be present in the clinical narrative, not merely asserted.
What's the difference between 99202, 99203, and 99204?
CPT 99202 requires straightforward MDM or 15 to 29 minutes of total time, 99203 requires low-complexity MDM or 30 to 44 minutes, and 99204 applies when a visit involves moderate-complexity MDM or reaches 45 minutes or more. A new patient presenting with two or more chronic conditions requiring active management, or a problem carrying moderate risk, shifts the encounter toward 99204.
What are the most common billing errors practices make with 99203?
Two of the most common errors are billing 99203 for established patients, by failing to verify whether any provider in the same specialty within the group saw the patient in the past three years, and underbilling by defaulting to 99203 when the clinical picture has already crossed into 99204 territory. Documentation failures, such as notes that assert low MDM complexity without the supporting clinical narrative, are another major source of miscoding.
Does a patient who sees a different doctor in the same group practice count as a new patient?
Not necessarily. Under AMA CPT Guidelines, a patient is only considered new if they have not received professional services from any physician of the same specialty and subspecialty within the same group practice in the past three years. The three-year rule applies across all physicians sharing the same group Tax Identification Number and specialty, so a patient returning to a different provider within the group before the three-year window has closed does not qualify as a new patient.
