46 New Patient CPT Codes: The Complete Reference Guide

A low denial rate is not proof your new patient coding is right. Here is where the revenue disappears silently, and what the audit trail never shows you.

The common assumption among practice administrators and billing decision-makers is that a low denial rate means the practice is coding new patient visits correctly and capturing its full earned revenue. A clean AR report feels like proof the billing is working. Claims go out, payments come back, denials stay low. For most practice administrators, that picture reads as confirmation that new patient coding is accurate enough. It is not.

"Billing departments are providing patients with incorrect or irrelevant CPT codes (e.g. 95165 for a new patient visit that involves no allergy shots or antigen mixing), creating confusion and potential billing errors."

— what we hear from medical billing professionals

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

The real problem is structural. Undercoded new patient visits pay out at the lower level without generating a denial, a flag, or any signal that revenue was left on the table. Standard AR aging reports track what was billed and whether it was paid. They cannot show what should have been billed but was not. That gap is invisible by design, and it compounds across every new patient encounter the practice sees. New patient office visits are billed on a four-code scale: 99202 through 99205, ascending by medical complexity and clinical work.

The Medicare reimbursement difference between 99203 and 99204 runs roughly $30 to $45 per encounter. A practice seeing 200 new patients per quarter and billing 80% of those at 99203 when documentation supports 99204 is forfeiting somewhere between $4,800 and $7,200 every three months, from a single code-level gap, with no claim ever rejected.

$7,200

Quarterly revenue lost from one code-level gap

Across the market, claims coded at a lower complexity level than warranted are paid at that lower level and do not generate denials or alerts, making revenue loss from undercoding invisible in standard accounts-receivable reporting. A claim paid at 99203 looks identical to a correctly coded 99203 claim.

The report has no way to distinguish them. The only diagnostic that surfaces this pattern is a coding distribution audit, which compares the practice's actual code frequency against documentation complexity and specialty benchmarks. According to PractiSynergy (2024), the overall accuracy rate for medical coding is approximately 55%, meaning nearly half of all medical claims contain some form of coding error.

That benchmark is the industry average. iScribe Health's Automated E&M Coding reads the clinical note in real time and surfaces the defensible code the documentation already justifies, an approach consistent with findings from Practolytics, which notes that automated tools analyzing documentation in real time generate code suggestions reflecting actual encounter complexity, helping convert MDM complexity that might otherwise go unrecorded into a recoverable line item before claim submission.

Key takeaways

• A low denial rate is not proof of accurate coding, it's proof that payers accepted what was submitted, which is a different thing entirely.
• Only ~55% of new patient E&M visits are coded at the correct level; the miscoding almost always runs downward, meaning practices are routinely billing less than the visit actually supports.
• CPT code 99201 was deleted in 2021, practices still running pre-2021 superbills are submitting invalid claims on a live billing cycle.
• The 3-year rule for new patient status is a three-part test, not a single date lookup; most billing teams check one condition and stop.
• Each code in the 99202-99205 range is governed by up to nine independently evaluated sub-criteria across two selection pathways, one underdocumented sub-element quietly pulls the whole code down.
• Code training solves the knowledge problem; it does not solve the timing problem, by the time a coder reviews a compressed note, the clinical complexity that warranted a 99204 or 99205 has already been flattened out of the record.
• iScribe Health's Automated E&M Coding generates E&M codes directly from clinical documentation in real time, capturing the complexity the visit actually reflects before the note is ever finalized.

How a New Patient Is Defined for CPT Coding Purposes Including the 3-Year Rule

Three tests stand between a patient and a new-patient code, and each one must clear before the next applies.

Skipping any layer means the status you assign may be wrong, and wrong status classifications on new-patient claims are among the most common triggers for post-audit takebacks, precisely because payers often pay them without flagging the error.

The AMA's Three-Part Test - Time, Physician, and Group Specialty Must All Clear

A new patient is one who has not received any professional services from the physician, or from another physician of the same specialty and subspecialty who belongs to the same group practice, within the past 3 years. That definition rests on three distinct elements:

• Time window, 36 months
• Physician identity test, same individual or same group
• Specialty and subspecialty match, both must align

All three must be evaluated on every encounter. This gap in the check is where overcoding silently accumulates: claims that are paid without denial, but that are wrong nonetheless.

Where the Group Practice Boundary Sits and the Specialty Exception That Changes Everything

Consider a patient who saw Dr. Smith (orthopedic surgery) in your group 18 months ago and now schedules with Dr. Jones (neurology) in the same group.

A patient who transitions from one specialty to a different specialty within the same group practice may qualify as a new patient, because the 3-year rule applies only to physicians of the same specialty and subspecialty, per the AMA CPT Guidelines. The table below summarizes how that boundary applies in practice.

• Scenario

Patient Status

• Same group, same specialty

Established, full stop

• Same group, different specialty

Potentially new

Key takeaway: A clean claim can be a wrong claim. Payers adjudicate and pay miscoded new-patient claims without flagging them; the denial rate measures whether a claim was processable, not whether the patient status was correct.

A clean claim can be a wrong claim.

iScribe Health's E&M Coding Intelligence, integrated directly into supported EHRs, is designed to evaluate clinical documentation at the point of encounter, specifically at the point of note completion, after the AI drafts the encounter summary, surfacing the patient-status determination before the claim is submitted.

Because iScribe Health's Ambient AI Documentation works through EHR Integration, physicians can keep their attention on the patient rather than the screen, restoring eye contact and the quality of the clinical encounter itself.

Why Prior Consultations Count Against the New Patient 3-Year Window

Prior consultations rendered by a physician of the same specialty within the same group count as professional services under the AMA definition, so they start the 3-year clock just as a standard office visit would. The two service types must be distinguished carefully, because receiving a consultation does not automatically substitute for or reset the new-patient designation under the 3-year rule.

Verifying whether a prior consultation exists in the patient's history before assigning new-patient status is therefore a necessary step in the workflow, and that is the operational gap iScribe Health's E&M Coding Intelligence is built to close.

Once patient status is confirmed, the next decision is which code in the 99202-99205 range applies. Before complexity levels can be mapped, there is a ghost in the code set that still derails billing teams: CPT code 99201, deleted in 2021, continues to generate claim errors in practices that have not fully removed it from their workflows.

What Happened to CPT Code 99201 and Why It Still Trips Up Billing Teams

Practices that last audited their superbills before that date may be sitting on a deleted CPT code that generates invalid claims every time a biller selects it without realizing anything is wrong, a silent leak that compounds across every patient encounter and every day of clinical practice.

Why Deleted Codes Do Not Disappear From Your Billing Workflow on Their Own

For clinical informatics teams already managing peak census periods, chasing down denial root causes that trace back to a deprecated template field is exactly the kind of administrative drag that pulls attention away from documentation quality. iScribe Health's Real-Time Denial Alerts and E&M Coding Intelligence are built for practices running supported EHRs that need a seamless layer of protection against this class of error, surfacing the problem at the point of note completion, before an invalid code ever reaches a payer.

The Floor Shifted to 99202 and Practices That Missed the Memo Are Miscoding the Lowest-Acuity New Patient Visits

Practices that missed this are miscoding the lowest-acuity encounters in their entire new patient population, an error that compounds silently in high-volume practices where clinicians are already charting two or more hours outside of patient care time and where no one has bandwidth to run a retrospective code audit.

Key takeaway: The deletion of that same figure is the clearest proof that the new patient code set demands active maintenance, not passive trust, and that maintenance cannot rely on care teams who are already stretched thin to catch it manually.

iScribe Health's Automated E&M Coding, delivered through EHR Integration with the platforms practices already run, is designed precisely to close this gap.

46 New Patient CPT Codes and Billing Rules Every Practice Needs to Know

The four new patient E/M codes look simple on the surface: match the visit to a level, submit the claim, move on. The problem is that each code is governed by up to nine discrete, independently evaluated sub-criteria across two selection pathways, and a single underdocumented sub-element silently pulls the entire code down. A claim that clears a payer's front-end edits can still represent thousands of dollars in annual undercoding that no denial report will ever surface.

The table below anchors the full reference that follows.

• CPT Code

MDM Level

Time Threshold

2026 Medicare Non-Facility Rate

• 99202

Straightforward

15-29 minutes

Lower anchor of the new patient range

• 99203

Low

30-44 minutes

Moderate step up from 99202

• 99204

Moderate

45-59 minutes

Frequently the highest-volume new patient code in primary care settings

• 99205

High

60-74 minutes

Highest reimbursement tier in the new patient range (verify current Medicare non-facility rate against the applicable CMS Physician Fee Schedule)

Highest reimbursement tier in the new patient range (verify current Medicare non-facility rate against the applicable CMS Physician Fee Schedule)

Code selection is based on either Medical Decision Making (MDM) complexity or total time on the date of encounter, whichever the provider chooses to document. The 40 reference items below cover every dimension of that choice.

1. CPT 99202 - New Patient, Straightforward MDM, 15-29 Minutes

99202 is the floor of the new patient range and the direct replacement for the deleted 99201. It applies when the encounter involves straightforward MDM (one self-limited or minor problem, minimal data, minimal risk) or when total provider time on the date of service falls between 15 and 29 minutes. Practices that default here out of caution when documentation supports that same figure or 99204 are leaving measurable reimbursement on the table with every encounter.

One reality that surfaces repeatedly in high-volume primary care settings: individual physicians and advanced practice providers are charting two or more hours outside of patient care time, compressing clinical detail into templated notes that simply cannot support codes above that same figure or 99203—not because the clinical work doesn't exist, but because the documentation doesn't capture it. Increasing patient throughput without adding headcount requires that every billable minute of clinical complexity makes it into the note, not just the diagnosis.

2. Medical Decision Making (MDM) Pathway - When to Use It Over Time

The MDM pathway is the preferred selection method when the clinical complexity of the visit is clearly documentable across all three sub-elements: problem complexity, data reviewed, and risk. It is especially advantageous for specialists whose visits are clinically dense but time-efficient. The tradeoff is that all three MDM sub-elements must independently meet the same level, a single weak sub-element pulls the entire code down, making careful documentation essential.

3. Total Time Pathway - When to Use It Over MDM

The time pathway is most practical when an encounter involves extensive counseling, shared decision-making, or care coordination that inflates total provider time beyond what MDM alone would support. It is also useful when MDM documentation is incomplete but time is well-documented. The trade-off: if a provider selects time as the basis for coding, every minute claimed must be defensible and tied to encounter-specific work, not estimated or rounded to a convenient threshold.

4. Total Time Definition - Face-to-Face Plus Non-Face-to-Face Work

Total time for E/M coding purposes includes all provider time on the date of the encounter: reviewing records before the visit, face-to-face time with the patient, ordering tests, documenting the note, and coordinating care after the patient leaves. This is not the same as appointment length. A 30-minute appointment slot can include 15 minutes of pre-visit chart review that legitimately pushes total time into 99204 territory, provided the note captures that work explicitly.

5. MDM Problem Complexity Sub-Element - What Counts at Each Level

Problem complexity is the first of three MDM sub-elements and the one most providers intuitively understand. It ranges from a single self-limited condition (straightforward, that same figure) through stable chronic illness (low, 99203), chronic illness with exacerbation or a new undiagnosed problem (moderate, 99204), up to a condition posing a threat to life or bodily function (high, 99205). The critical documentation requirement: the note must name the specific problem and its status, not just list diagnoses without clinical context.

6. MDM Data Reviewed Sub-Element - Tests, Records, and Independent Interpretation

The data sub-element trips up more practices than any other MDM component. At the moderate level (99204), the provider must meet a three-category threshold: any combination of reviewing external records, ordering or reviewing tests, or independently interpreting a test result. "Independent interpretation" means the provider personally reviewed and documented their own interpretation of an image, tracing, or specimen, not simply noting that a result was received. Missing this notation is one of the most common reasons a documented that same figure encounter gets pulled back to 99203.

This is exactly the kind of sub-element gap that iScribe Health's E&M Coding Intelligence is built to catch. At the point of note completion, after the AI drafts the encounter summary, the system surfaces specific missing notations such as the absence of an independent interpretation statement, before the claim is submitted. For practices where clinicians are regularly charting under time pressure, that real-time signal is the difference between a defensible that same figure and a quietly downgraded 99203 on every qualifying encounter.

7. MDM Risk Sub-Element - Prescription Drug Management as Moderate Risk

The result is documentation that reflects the actual risk sub-element, not a redacted version of it.

8. Post-2021 Rule - History and Exam No Longer Drive Code Selection

Since the that same figure AMA E/M guideline revisions, history and physical examination are no longer the primary drivers for selecting a new patient office visit level. Code selection is now based on MDM complexity or total time on the date of encounter. This is the single most consequential rule change in recent E/M history. Practices whose documentation workflows were built around capturing history elements and exam bullet points are structurally producing notes that cannot support codes above that same figure, regardless of the actual clinical complexity delivered.

9. Medically Appropriate History Requirement - What Documentation Satisfies It

History is still required, but the standard shifted from prescriptive (chief complaint, HPI elements, ROS, PFSH) to clinically appropriate. The provider documents the history that is relevant to the presenting problem and the clinical decision being made. A focused history for a straightforward acute complaint satisfies this requirement just as fully as an extensive history for a complex new patient, as long as the note reflects what was medically necessary for that encounter. The key is that the history documented supports the clinical decisions recorded in the assessment and plan.

10. Medically Appropriate Examination Requirement - Scope and Documentation

The same principle applies to the physical examination: document what is medically appropriate for the presenting problem, not a fixed number of organ systems or body areas. The that same figure overhaul eliminated the multi-system exam bullet-counting approach entirely. For audit purposes, the examination documented must be consistent with the problem complexity claimed in the MDM. A note claiming high-complexity MDM for a systemic condition while documenting a single-system exam will draw scrutiny from reviewers looking for internal consistency across the note.

11. New Patient Definition - The 3-Year Rule for Code Eligibility

A patient qualifies as new only if no professional services have been rendered by the physician, or by any physician of the same specialty and subspecialty in the same group practice, within the past three years. This rule has meaningful group-practice implications: a patient who saw a cardiologist in the group two years ago and now presents to a different cardiologist in the same group is an established patient, not a new one. Billing a new patient code in that scenario creates an overpayment exposure that payers can recover retrospectively.

It is also worth noting that patients themselves are frequently confused about what a new patient visit code represents, particularly when they believe a visit was routine or vaccine-only and are then billed a that same figure. That confusion is a compliance signal as much as a patient-satisfaction issue. When documentation is thorough and the note clearly reflects the clinical assessment that justified the code, practices have a defensible answer to those inquiries. Standardizing clinical documentation quality across the practice is the structural fix; iScribe Health's AI Customization layer ensures that documentation standards are applied consistently at the individual provider level, not left to variation in charting habits.

12. Prolonged Service Add-On Code for New Patients Exceeding 99205 Time Ceiling

When total provider time on the date of service exceeds 74 minutes for a new patient visit, the base code 99205 is billed with add-on code 99417 for each additional full 15 minutes beyond the threshold, for non-Medicare payers. 99417 cannot be billed in partial increments; only completed 15-minute blocks count. This code applies to new patients specifically when using the 99205 time pathway. Practices that do not bill 99417 on qualifying encounters are systematically undercoding their longest, most complex new patient visits.

13. Medicare G2212 - Prolonged Service Add-On for Medicare New Patient Visits

Medicare does not recognize that same figure for prolonged services. For Medicare new patient visits exceeding the that same figure time ceiling, the correct add-on is G2212, which applies after 74 minutes of total provider time and is billed in full 15-minute increments beyond that threshold. Using that same figure on a Medicare claim will result in a denial or non-payment.

Commercial payers vary: some follow the G2212 convention, others accept that same figure. Verify each payer's policy before billing prolonged service add-ons to avoid systematic claim errors. iScribe Health's Real-Time Denial Alerts flag these payer-specific mismatches before they reach the denial queue, which is particularly valuable in multi-payer environments where the that same figure vs.

G2212 distinction is a recurring source of claim errors.

14. Incident-To Billing Restriction - New Patients Cannot Be Billed Incident-To

Incident-to billing, where a non-physician practitioner's (NPP) services are billed under the supervising physician's NPI at the full physician rate, is not permitted for new patient visits. The incident-to rules require that the physician personally initiate the plan of care for the patient, which by definition has not occurred with a new patient. Billing a new patient NPP encounter incident-to the physician is a compliance exposure. The NPP's services must be billed under the NPP's own NPI, typically at 85% of the physician fee schedule rate.

15. Medicare Annual Wellness Visit - Distinct from New Patient E/M Codes

The Medicare Annual Wellness Visit (AWV) is a preventive benefit, not an E/M service, and is billed under G0438 (initial AWV) or G0439 (subsequent AWV), not under the 99202-99205 range. If a new patient's first Medicare visit involves both a preventive AWV and a significant, separately identifiable E/M service, the E/M code may be billed alongside the AWV with modifier 25 appended. Conflating these two service types or billing a new patient E/M instead of the AWV when the visit was purely preventive creates a coding mismatch that payers flag on review.

16. Medicare Advantage Plan Variations - Payer-Specific New Patient Billing Rules

Medicare Advantage (MA) plans are required to cover the same benefits as traditional Medicare, but individual plans may impose additional documentation requirements, prior authorization rules, or coding policies that differ from CMS fee-for-service guidelines. Some MA plans require specific clinical data elements in the note to support that same figure or 99205. Practices billing new patient visits across multiple MA plans should maintain a payer-specific reference for documentation requirements, because a note that satisfies CMS standards may still be downcoded by an MA plan's internal review criteria.

17. MDM Straightforward - 99202 Problem Complexity - One Self-Limited Condition

At the straightforward MDM level, problem complexity is defined as a single self-limited or minor problem. Classic examples include an acute upper respiratory infection, a minor skin abrasion, or a single-episode headache with no red flags. The note must name the specific condition and reflect that the clinical assessment concluded the problem was self-limited. Providers who document "patient presents with cold symptoms, treated symptomatically" without naming the diagnosis and its status are creating ambiguity that reviewers can use to question even a that same figure claim.

18. MDM Low - 99203 Problem Complexity - Two Self-Limited or One Stable Chronic Condition

Low MDM problem complexity covers two or more self-limited problems addressed in the same encounter, or one stable chronic illness being managed. A stable chronic illness means the condition is being managed as expected, with no exacerbation, progression, or side effects requiring management changes. The documentation must reflect stability explicitly, not just list the chronic condition. A note that lists a chronic condition without documenting current clinical findings, adherence, and status does not adequately support the "stable" characterization for audit purposes.

19. MDM Moderate - 99204 Problem Complexity - Chronic Illness with Exacerbation or New Undiagnosed Problem

Moderate MDM problem complexity, which anchors that same figure, covers a chronic illness with exacerbation, progression, or side effects of treatment; a new problem with uncertain prognosis; or an acute illness requiring prescription drug management. that same figure frequently represents the highest-volume new patient code in primary care settings and reflects the typical scenario involving multiple conditions or a condition requiring active management decisions. The documentation must capture what makes the condition non-stable or the prognosis uncertain, not simply list the diagnosis.

20. MDM High - 99205 Problem Complexity - Threat to Life or Bodily Function

High MDM problem complexity, which supports that same figure, requires a severe exacerbation of a chronic illness, a complex chronic condition requiring intensive management, or an acute or chronic illness posing a threat to life or bodily function. The Medicare non-facility rate for that same figure is the highest in the new patient range; practices should verify the current figure against the applicable CMS Physician Fee Schedule before relying on it for revenue modeling. The documentation must reflect the severity and the clinical reasoning behind the high-complexity assessment. A note that lists a serious diagnosis without documenting the management complexity will not survive audit scrutiny at this level.

21. MDM Low Data - 99203 Data Sub-Element - Minimal Review or Ordering

At the low MDM data level, the provider must meet a minimal threshold: ordering or reviewing one test, or reviewing external records, or independently interpreting a single result. This is a low bar, but it still requires explicit documentation. Providers who review prior records or order labs during a that same figure encounter but do not document that review in the note are failing to capture a sub-element that is already present in the clinical work. The gap is not in the care delivered; it is in the documentation that records it.

22. MDM Moderate Data - 99204 Data Sub-Element - Three-Category Threshold

Moderate MDM data requires meeting criteria in at least three of the following categories: reviewing external records or ordering tests (each counts separately), independently interpreting a test result, or discussing the case with an external physician or appropriate source. This three-category threshold is where most that same figure documentation failures occur. A provider who reviews outside records and orders two tests has met the threshold but must document each action specifically. Generic phrases like "reviewed prior records" without identifying what was reviewed and what clinical impact the review had are insufficient for a defensible that same figure.

iScribe Health's Ambient Listening and Conversational AI captures these data-review discussions as they happen during the encounter, when the provider references outside records or discusses a test result aloud, so the documentation reflects each category specifically rather than collapsing into a generic phrase post-visit.

23. MDM High Data - 99205 Data Sub-Element - Two or More High-Complexity Data Elements

High MDM data requires meeting criteria in at least two of the following: reviewing and summarizing external records from multiple sources, independently interpreting a test result with documented findings, or discussing the case with an external physician or appropriate source where that discussion influenced management. The documentation standard at this level is explicit and specific. Practices whose notes contain high-complexity clinical work but use templated, generic data review language are producing notes that cannot support that same figure regardless of what actually occurred in the encounter.

24. MDM Low Risk - 99203 Risk Sub-Element - OTC Medications or Minor Surgery

Low risk in the MDM framework covers management decisions involving over-the-counter medications, minor surgery with no identified risk factors, or physical therapy referrals. This sub-element is often the easiest to satisfy at the that same figure level but also the easiest to leave undocumented. A provider who recommends ibuprofen for a new patient's acute pain must document that recommendation as a management decision, not just list it in the plan section without clinical context. The risk sub-element requires evidence of a decision, not just a notation of what was prescribed.

25. MDM High Risk - 99205 Risk Sub-Element - Drug Therapy Requiring Intensive Monitoring

High risk covers management decisions involving drug therapy requiring intensive monitoring for toxicity, decision-making under uncertainty with significant risk of morbidity, or hospitalization. Initiating anticoagulation therapy, managing a patient on chemotherapy, or making a decision to pursue or withhold aggressive intervention all qualify. The documentation must reflect the clinical reasoning behind the high-risk decision, not just the decision itself. A note that records the initiation of a high-risk medication without documenting why, the monitoring plan, and the risk-benefit discussion does not satisfy this sub-element for audit review.

26. Time Documentation Best Practice - Recording Start, Stop, or Total Minutes

When using the time pathway, the note must document either the start and stop time of all provider work on the date of service, or the total number of minutes spent. Both approaches are acceptable; the key is that the time documented is specific, plausible, and consistent with the clinical work described elsewhere in the note. Documenting "approximately 45 minutes" without specificity invites payer scrutiny. Documenting a precise total that is inconsistent with the complexity of the note creates a different kind of audit risk. Specificity and internal consistency are both required.

27. Pre-Visit Work Included in Total Time - Chart Review Before the Encounter

Chart review performed before the patient encounter on the same date of service counts toward total time for E/M coding purposes. This includes reviewing prior visit notes, imaging reports, specialist correspondence, and outside records. The documentation must explicitly state that pre-visit chart review occurred and, ideally, identify what was reviewed. Practices whose clinical documentation workflows capture only in-room encounter elements are structurally incapable of supporting codes above that same figure for time-based billing, because the pre-visit work that would push total time into 99204 or 99205 territory is never recorded.

28. Post-Visit Work Included in Total Time - Documentation and Care Coordination

Time spent after the patient leaves, including completing the clinical note, reviewing results that return the same day, writing referral letters, and coordinating with other providers, counts toward total time on the date of the encounter. This post-visit work is frequently the difference between a that same figure and a 99204 time threshold. The challenge is that most EHR workflows do not prompt providers to document post-visit time separately from the note itself.

Without a deliberate documentation habit or a tool that captures this work, the time is performed but never billed. iScribe Health's EHR Integration surfaces this gap within the existing EHR workflow, without requiring providers to switch platforms or adopt a parallel documentation system, so that post-visit work recorded during note completion is available to the E&M Coding Intelligence layer at the moment the code is assigned.

29. Counseling Time Included in Total Time - Patient Education and Shared Decision-Making

Time spent on patient education, discussing diagnosis and prognosis, reviewing treatment options, and facilitating shared decision-making all count toward total time on the date of service. This is particularly relevant for new patient visits where providers spend significant time explaining a new diagnosis or a complex management plan. The note should document what was discussed, with whom, and for approximately how long. Counseling time that is performed but not documented is invisible to the coding process and cannot be used to support a higher-level time-based code.

30. Care Coordination Time - Calling Specialists or Arranging Referrals on Encounter Date

Care coordination performed on the same date as the encounter counts toward total time, including time spent calling specialists, arranging referrals, coordinating with pharmacies, or communicating with other members of the care team about the patient's management plan. This work often occurs after the patient has left the office and is frequently undocumented. For new patients with complex needs, care coordination time can add 15 to 20 minutes to total encounter time, potentially moving the code from that same figure to 99204, but only if the note captures that work explicitly.

31. Split/Shared Visit Rules for New Patients - Physician vs. NPP Billing

A split/shared visit occurs when both a physician and a non-physician practitioner (NPP) provide substantive portions of an E/M service on the same date. For new patient visits, the visit is billed under the provider who performs the substantive portion of the encounter, defined since 2022 as the provider who performs and documents the history, physical exam, or medical decision-making. If the physician performs the MDM and the NPP performs the history and exam, the physician's role in MDM determines billing rights. Documentation must clearly attribute each clinical element to the specific provider who performed it.

32. Same-Day New Patient Visit and Procedure - Modifier 25 Requirement

When a new patient visit and a procedure are performed on the same date of service, the E/M service must be billed with modifier 25 to indicate that it is a significant, separately identifiable evaluation and management service beyond the pre-procedure assessment included in the procedure's global package. Without modifier 25, payers will bundle the E/M into the procedure payment and deny the E/M charge. The E/M note must document a clinical assessment and management decision that is distinct from the procedure itself, not simply a note that the procedure was planned.

33. Global Surgery Period - New Patient E/M During Postoperative Period

If a new patient presents during the global surgery period of a procedure performed by the same surgeon, the E/M service is generally included in the global package payment and cannot be billed separately unless the visit is for a problem unrelated to the surgical procedure. When billing a new patient E/M during a global period for an unrelated problem, modifier 24 is required. Documentation must clearly establish that the presenting problem is unrelated to the surgical condition. Practices that miss this distinction either fail to bill legitimate E/M services or create overpayment exposure by billing without the required modifier.

34. Specialty-Specific New Patient Coding - Psychiatry and Behavioral Health Nuances

Psychiatry and behavioral health practices use the standard that same figure-99205 range for new patient office visits but frequently bill 99204 as the baseline for initial psychiatric evaluations, given the inherent complexity of new psychiatric assessments involving multiple conditions, medication history review, and risk stratification. The MDM data sub-element is particularly relevant in psychiatry: reviewing prior psychiatric records, coordinating with other treating providers, and independently assessing prior diagnoses all contribute to the three-category moderate data threshold. Practices in this specialty that default to that same figure for initial evaluations are systematically undercoding their most complex new patient encounters.

35. Consultation Codes vs. New Patient Codes - When Each Applies

Consultation codes (99241-99245 for office consultations) require three specific elements: a request from another provider, the rendering provider's opinion or advice, and a written report back to the requesting provider. Medicare stopped recognizing consultation codes in 2010 and requires providers to bill new patient or established patient E/M codes instead. Commercial payers vary: some still recognize consultation codes, others follow the Medicare model. Practices that bill consultation codes to payers that do not recognize them, or that bill new patient codes when consultation codes are appropriate and recognized, create both denial exposure and potential undercoding.

36. Documentation Anchor - Assessment and Plan as the MDM Centerpiece

The assessment and plan section of the clinical note is the primary documentation anchor for MDM-based coding. This is where problem complexity, data reviewed, and management risk must be explicitly captured. A note with a thorough history and examination but a thin assessment and plan cannot support a high-level MDM code, regardless of the clinical work performed. The assessment must name each problem addressed, its status, and the clinical reasoning behind the management decision. The plan must document specific actions taken, referrals made, and follow-up instructions, all tied to the problems identified.

37. EHR Template Risk - Auto-Populated Fields and Audit Vulnerability

EHR templates that auto-populate history, review of systems, or physical examination fields create significant audit vulnerability for new patient visits. When identical or near-identical language appears across multiple new patient notes, payers and OIG auditors flag the pattern as cloned documentation, which raises questions about whether the documented elements actually occurred. The OIG has identified cloned documentation as a driver of E/M overpayment recoveries in audit findings targeting new patient visits. Practices that rely on auto-populated templates without provider-specific, encounter-specific customization are building audit exposure into every note that template generates.

iScribe Health's AI Customization layer is designed specifically to counter this pattern. Rather than pre-filling fixed template language, the Ambient Listening and Conversational AI generates encounter-specific documentation from the actual clinical conversation, producing notes that are differentiated by encounter content, not by template slot. For health systems where standardizing clinical documentation quality across the practice is a priority, this means consistency in structure and MDM capture without the cloning risk that uniform templates create.

38. Payer-Specific Documentation Requirements - Commercial Insurers vs. Medicare

While Medicare and most Medicaid programs follow the 2021 AMA E/M guidelines for new patient code selection, many commercial payers have not uniformly adopted these standards. Some commercial insurers still require documentation of history elements, review of systems, or a specified number of exam systems for higher-level new patient codes. Practices must maintain payer-specific documentation policies and train providers accordingly, as a note that satisfies Medicare may be insufficient for a commercial payer audit.

39. Upcoding Risk - Statistical Outlier Monitoring for New Patient Code Distribution

CMS and commercial payers monitor E/M code distribution by specialty and flag providers whose billing patterns fall outside expected norms. A primary care practice billing that same figure for 60% of new patient visits will draw scrutiny regardless of documentation quality, because that distribution is statistically inconsistent with specialty benchmarks. The risk is not just audit exposure; it is that the resulting review will examine every claim in the outlier range. Practices should benchmark their own new patient code distribution against specialty norms before assuming that higher-level codes are always the right target, even when documentation supports them.

40. Undercoding Risk - Revenue Loss from Defaulting to Lower New Patient Codes

Undercoding is the more common and less visible problem for most practices. Providers who default to that same figure or 99203 out of audit fear, or because documentation does not explicitly capture MDM elements, are systematically billing below the level the clinical work supports. Because undercoded claims pay without triggering a denial, the revenue loss is invisible in standard AR reports. The gap only surfaces when coding distribution is audited against documentation complexity. Most billing teams assume their coding is correct because claims are paying; as the revenue-gap analysis in the next section shows, that assumption is the most expensive one a practice can make.

Most billing teams handle this by relying on provider judgment at the point of coding, trusting that if the claim paid, the level was defensible. The hidden cost is that the note created under time pressure rarely captures the MDM sub-elements that would support a higher-level code, so the provider's clinical judgment about complexity never makes it into a billable, defensible record.

The tool materializes its full value when the practice or health system is already running a supported EHR and wants a seamless ambient documentation experience, with iScribe Health's EHR Integration handling the connection so that the workflow change for the individual physician or advanced practice provider is minimal. Documentation is captured during or immediately after the encounter, where the clinical detail is still present and can be recorded before it is compressed into a templated note.

Having every rule in this reference is a necessary starting point, but it raises a harder question: if practices already have access to the code definitions, why does the revenue gap persist at scale? The next section delivers the answer, and it has less to do with what billing teams know than with what clinical documentation consistently fails to capture.

Related Reading

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

Why Knowing Every New Patient CPT Code Still Leaves a Measurable Revenue Gap

Key takeaway: Only 54.8% of new patient E&M visits were coded at the correct level in a real-world orthopedic audit of 941 encounters (Orthobullets Evidence, PMID 32379124), so the other 45% carried active compliance and revenue risk, regardless of whether a single denial had been filed.

What a 941-Encounter Audit Reveals About the 55% E&M Accuracy Baseline

According to Orthobullets Evidence (PMID 32379124), providers and an AI coding tool agreed on the E&M level in only 54.8% of 941 orthopedic encounters, so nearly half of all visits carried a discrepancy between the documented code and the warranted level. The same orthopedic audit found a 33% overcoding rate within that specialty sample, a figure that surfaces a counterintuitive risk: overcoded claims often pay without dispute, masking the compliance exposure entirely until a payer initiates a retrospective review. As Becker's Hospital Review has documented, the complexity of orthopedic encounters routinely outpaces what standardized codes are designed to capture, compounding both the documentation burden and the coding accuracy problem.

Why Time-Pressured Notes Silently Downgrade Defensible High-Level Codes

As Orthobullets Evidence (PMID 32379124) makes clear, the operative variable is whether the documentation created at the time of service can actually support the level selected, and under time pressure that standard is routinely missed even when the clinical complexity was genuinely present.

The root cause of this collapse is time pressure at the point of documentation. iScribe Health's Ambient AI Documentation, built on Ambient Listening and Conversational AI, accurately captures clinical notes during or after patient encounters, reducing the time physicians spend on documentation so more of their attention stays on patient care rather than on reconstructing encounter details hours later. Downstream, iScribe Health's Automated E&M Coding and Real-Time Denial Alerts ensure that what reaches the payer is both accurate and compliant, ensuring accurate, compliant medical coding across high patient volumes to maximize reimbursement and minimize claim denials.

The Coding Distribution Blindspot That Turns One-Time Errors Into Compounding Revenue Loss

Most AR dashboards surface denials and underpayments but offer no visibility into whether the codes submitted in the first place reflected the complexity of the encounters they represented.

Because iScribe Health integrates directly with supported EHRs, the ambient documentation and E&M coding intelligence operate within the existing clinical workflow rather than adding a separate tool layer.

How Automated E&M Coding Closes the New Patient Documentation Gap at Scale

It does not solve the timing problem.

No amount of code knowledge recovers what documentation never captured. As Capline Healthcare Management notes, when notes are incomplete or compressed, coders default to lower-level codes that do not reflect the actual work performed, resulting in systematic undercoding and lost revenue.

Why Real-Time MDM Capture Changes the Code Not Just the Note

Medical decision-making elements (problem complexity, data reviewed, and risk assessed) are the clinical facts that separate a that same figure from a 99204. The problem is that those facts exist at the moment of the encounter, not when the note gets finished two hours later. According to industry research, capturing MDM elements as documentation is created builds a defensible coding record for higher-complexity visits that would otherwise default to a lower-level code. Real-time capture does not change what the provider did. It changes whether the record reflects it.

Why Documentation Quality Is the Ceiling Automated Coding Cannot Raise Alone

Automated E&M coding tools are only as accurate as the documentation feeding them. A coding engine cannot surface MDM complexity that was never recorded. The familiar approach, having providers dictate or type notes after the visit, creates the compression that forces coders toward the safest defensible code. Practolytics is direct on this point: automated tools analyze clinical documentation in real time to generate code suggestions that reflect actual encounter complexity, removing reliance on a coder's interpretation of an incomplete note.

The accuracy data makes one conclusion unavoidable. The revenue gap is the statistical norm:

• [~the same pattern holds.com/ensuring-accurate-medical-coding-to-minimize-billing-errors-in-healthcare/)
• Provider, AI agreement at just 54.8% across 941 real-world orthopedic encounters

Practices that pair an AI medical scribe with automated E&M coding are positioned to close the timing gap at its source. Complexity is documented as it happens, MDM elements are surfaced before the note is finalized, and the resulting code can reflect what actually occurred in the room rather than what survived compression into an after-the-fact note.

The code knowledge is already in this guide.

Next steps

If your denial rate looks clean while your new patient claims quietly pay out at 99203 when documentation supports 99204, the path forward starts with closing the gap at the point of documentation, not in a retrospective coding review. Start with our AI medical scribe.

The 2021 E/M overhaul removed history and physical exam as code-level drivers, meaning any practice whose note-capture workflow misses pre-visit chart review, post-visit coordination, or prescription management decisions is structurally capped below the codes the clinical work actually justifies. That structural ceiling compounds into tens of thousands of dollars annually because undercoded claims pay without triggering a denial, making the revenue loss invisible in standard AR reports until a distribution audit surfaces it. With industry-wide E/M coding accuracy hovering around 55%, that audit almost always confirms the same finding: systematic undercoding of 99204 and 99205 encounters that paid, uncontested, at the wrong level.

Together, these two realities point to one intervention: capturing MDM complexity in real time, before the note is compressed and the code is locked.

Start with an AI medical scribe that generates encounter documentation during the visit and surfaces complexity-matched code suggestions at the point of note completion, before a single claim goes out. From there, coding distribution shifts to reflect what your clinical work actually supports, not what survived after-hours charting.

Frequently Asked Questions

How do I know whether to bill a returning patient as new or established?

You have to pass all three parts of the AMA's test: the patient must not have received any professional services within the past 36 months, from any physician of the same specialty and subspecialty, within the same group practice. If any one of those three conditions fails, time window, physician identity, or specialty match, the patient is established, not new.

If a patient saw a different doctor in our group, does that make them established?

It depends on specialty. If the prior physician was the same specialty and subspecialty within your group, the patient is established, full stop, even if your specific provider has never personally treated them. If the prior physician was a different specialty, the patient may still qualify as new, because the 3-year rule applies only to physicians of the same specialty and subspecialty per the AMA CPT Guidelines.

Does a prior consultation count as a professional service for the 3-year rule?

Yes. A prior consultation by a same-specialty physician in the same group still counts as a professional service for purposes of the time-window test, as set out in the AMA CPT Guidelines, and does not reset or bypass the new-patient designation.

We have a low denial rate, doesn't that mean our new patient coding is accurate?

Not necessarily. Undercoded claims are paid at the lower level without generating a denial, a flag, or any alert, so a clean AR report cannot show what should have been billed but wasn't. The only diagnostic that surfaces undercoding patterns is a coding distribution audit that compares actual code frequency against documentation complexity and specialty benchmarks.

Is CPT code 99201 still valid to bill for new patient visits?

No. CPT code 99201 was permanently deleted effective January 1, 2021, and any claim submitted with it after that date is invalid and deniable regardless of the accuracy of the clinical documentation. The current floor for new patient office visits is 99202, and any billing workflow or superbill still listing 99201 as a selectable option needs to be updated immediately.