Templates

CPT Code 99203 With Examples and the Billing Errors That Send It Back

CPT Code 99203 With Examples and the Billing Errors That Send It Back

CPT Code 99203 With Examples and the Billing Errors That Send It Back

1 min read

99203 is what a new patient visit gets coded when nobody checked.

It sits in the middle of the new patient range, which makes it feel like the cautious choice. It is not a cautious choice. It is a default, and a default is wrong in both directions: visits that were genuinely straightforward get coded 99203 when 99202 was right, and visits that reached moderate decision making get coded 99203 when 99204 was right.

Here is what the code is, the errors that send a claim back, where the line sits against the codes on either side, what it pays, and a documentation template with two worked examples.

Key Takeaway: CPT code 99203 is the office or other outpatient visit for a new patient requiring a medically appropriate history and examination with a low level of medical decision making, or 30 to 44 minutes of total time on the date of the encounter [1]. The errors that send it back are a patient who was not new, a note evidencing neither route, a time figure recorded without being qualified, a level chosen by habit rather than against the criteria, and a diagnosis that does not support medical necessity. Four of those five are settled before the claim is submitted. What it pays is not one number: Medicare computes the fee schedule amount as the relative value units (RVUs) for the service, multiplied by the geographic adjustment factor (GAF) for your fee schedule area, multiplied by the conversion factor (CF) [2], and a non-participating supplier is paid 95 percent of the participating amount [2]. Sully.ai drafts the note with the detail that supports the level you actually delivered.

What CPT Code 99203 Is

99203 is the office or other outpatient visit for a new patient. The descriptor asks for a medically appropriate history and examination, and a low level of medical decision making, or 30 to 44 minutes of total time on the date of the encounter [1].

It is the middle of three codes most practices use for new patients. 99202 sits below it at straightforward decision making. 99204 sits above it at moderate. All three are new patient codes, so the question of which series you are in is settled before any of this applies.

New Patient, and the Three Year Rule

A patient is new if they have not received a professional service in the past three years from a physician of the same specialty and subspecialty, in the same group practice.

The trap worth repeating: a colleague in your own specialty, in your group, makes the patient established with you even if you have never met them. Covering a partner's panel does not reset the clock. The rule is covered at length, with the other two conditions, in the 99204 article.

The Two Routes to 99203

You need one route, not both. Either 30 to 44 minutes of total time on the date of the encounter, or a low level of medical decision making [1][3].

Low decision making means two of the three elements reach low: one stable chronic illness or two or more self-limited problems under problems addressed, limited data reviewed, and a low risk of complications from management. The low end of that grid is laid out in the 99213 article, and the full framework, including what two of three means, is in the 99214 article.

Ladder of the three new patient office visit codes showing 99204 at moderate decision making or 45 to 59 minutes, 99203 at low or 30 to 44 minutes, and 99202 at straightforward or 15 to 29 minutes, with the single change that moves a visit between adjacent rungs

The Billing Errors That Send a 99203 Back

Five errors account for most of them. Four are settled before the claim is submitted, which means they are visible to the practice while they are still fixable.

The Patient Was Not New

Someone of your specialty and subspecialty, in your group practice, provided a professional service inside the past three years. The established patient code applies instead.

This is the error that produces the cleanest denial, because it is the one a payer can verify automatically against its own claims history. It does not require anyone to read the note.

The reverse error produces nothing at all. Billing an established patient code for a genuinely new patient generates no denial, no flag and no correspondence. It just pays less, on every misclassified visit, until somebody notices. Almost nobody notices.

The Note Evidences Neither Route

No total time recorded on the date of the encounter, and the note does not establish two of three decision making elements at low. Without one of the two, nothing supports the level.

The fix is a single line. State which route you used, explicitly, in the note. A reviewer should not have to work out which case you are making.

Time Was Recorded Without Being Qualified

This is the most specific error on the list and the least written about.

"35 minutes" is not a time entry. It is a number with no stated meaning, and a reviewer has no way to tell what it measures.

Two qualifications have to be visible. That the figure is total time on the date of the encounter, rather than face to face time, which is the thing most clinicians instinctively record. And that it counts only the reporting clinician's time [3]. A figure that includes a medical assistant's rooming time, a nurse's intake, or a scribe's documentation does not support the route, and the bare number conceals that.

Write it as "total time on the date of the encounter: 36 minutes", and say what it comprised.

The Default Error, Which Runs in Both Directions

The one that generates the fewest denials and the most inaccurate records.

99203 is the new patient default the way 99213 is the established patient default. It gets chosen because it is the middle of the range and nobody questions the middle.

Downward. A genuinely straightforward visit, one self-limited complaint, no data reviewed, minimal risk, coded 99203 on the reasoning that it was a new patient so it is at least a level three. That visit is a 99202 [3].

Upward. Two problems addressed, or data independently interpreted, or a new prescription started, coded 99203 because the visit felt routine and 99204 felt like a reach. That visit is a 99204.

To be clear, this is not an argument for coding up. It is an argument against coding by habit, which produces errors in whichever direction the habit happens to point. The correct code is the one the work supports, and reading the note against the criteria is the only way to find it.

The Diagnosis Does Not Support Medical Necessity

The only one on the list that is a clinical question rather than a process question. The visit can be documented perfectly and still be denied if the diagnosis submitted does not justify a new patient visit at that level.

The five billing errors for CPT 99203 placed at the stage where each is created, across scheduling, the visit, coding and submission, showing that four of the five are settled before the claim is submitted

99202, 99203 or 99204

All three are new patient visits, so the three year rule is already settled and the only question left is the level.

Code

Medical decision making

Total time on the date of the encounter

99202

Straightforward

15 to 29 minutes

99203

Low

30 to 44 minutes

99204

Moderate

45 to 59 minutes

Sources [1][3].

What puts a visit at 99202. One self-limited or minor problem. No data reviewed. Minimal risk from management. A new patient with a single uncomplicated complaint, examined, reassured, nothing prescribed.

What moves a visit to 99204. A second problem addressed rather than one. Data independently interpreted rather than noted. Prescription drug management, which reaches moderate risk on its own. Any two of those three put the visit above 99203, and the detail is in the 99204 article.

The time boundaries are one minute wide. A 29 minute visit is a 99202 on time. A 30 minute visit is a 99203. A 44 minute visit is a 99203. A 45 minute visit is a 99204. There is no rounding convention that makes "about half an hour" into a usable entry, which is the strongest practical argument for writing down the actual figure.

What 99203 Actually Pays

There is no single national figure, and any page quoting one is either out of date or correct for exactly one part of the country.

Under 42 CFR 414.20, the fee schedule amount for a participating supplier is computed as the RVUs for the service, multiplied by the GAF for the fee schedule area, multiplied by the CF [2]. Two of the three move: the geographic adjustment factor differs in every fee schedule area, and the conversion factor is set annually and revised most years.

RVUs are established separately for physician work, practice expense and malpractice insurance [4], and the practice expense component is lower when you provide the service in a facility than in an office, because the facility carries the overhead. One rule that routinely gets missed: a non-participating supplier is paid 95 percent of the participating amount [2].

Get your own figure from the CMS Physician Fee Schedule Look-Up Tool [5]. Four inputs: the year you are billing for, the code, your Medicare Administrative Contractor (MAC) locality, and facility or non-facility. Commercial payers set their own rates, frequently as a percentage of the Medicare schedule, so treat the Medicare figure as your reference point rather than your full picture.

If you are an employed physician asking what the visit is worth to you specifically, that is a different question with a different answer. The work RVU, and why it does not move with your locality, is covered in the 99204 article.

The 99203 Documentation Template

One block, complete as written. Paste it straight into a chart, or start from the progress note template and add the new patient lines.

OFFICE VISIT, NEW PATIENT

  • Patient, date of service: (and the rendering clinician)

  • New patient confirmation: (no professional service from a same-specialty clinician in this group within three years, with the date checked)

  • Referral source: (who sent them, and the question being asked)

  • Chief complaint: "(in the patient's words)"

  • History: (the full history, since none of it is on file yet)

  • Examination: (findings relevant to the problems being addressed, with values)

  • Problems addressed today: (each one named, with its status: stable, worsening, new, self-limited)

  • Data reviewed: (outside records, labs, imaging, notes from other clinicians, or "none" if none)

  • Risk of management selected: (what you decided, and what the risk of that decision is)

  • Route used: (time, or medical decision making; state which)

  • If time: (total time on the date of the encounter, stated in those words, counting only this clinician, and what it comprised)

  • Assessment and plan, per problem: (each problem gets its own assessment and its own plan, the same structure a SOAP note uses)

  • Signature, credentials, date: (the clinician who performed the service)

Three lines carry the level. "New patient confirmation" decides the code series. "Problems addressed today" and "Risk of management selected" decide the decision making route.

The "If time" line is written the way it is on purpose. It is the only line in this template built to prevent a specific denial, and the phrasing is the fix.

Two parallel checklists of what a note must carry for each route to 99203, the total time route and the medical decision making route, with the most commonly missing lines highlighted

Where Sully.ai Fits in a New Patient Visit

A first visit carries a history that has never been taken, outside records read for the first time, and reasoning spoken aloud in the room that has to survive into a note written three patients later.

What gets lost is the detail that evidences the level, in whichever direction the level actually sits. The second problem you addressed briefly. The record you read before walking in. Equally, the fact that there was only one problem and nothing was prescribed.

Sully.ai's AI scribe drafts the note from the visit, so the problems addressed, the data reviewed and the risk are on the page rather than reconstructed later. It runs on a single integration across Epic, Cerner, Meditech and Athenahealth.

Two boundaries, and both are refusals.

Nothing should choose the code. Code selection is your attestation, it is what an auditor examines, and it carries liability belonging to the person who saw the patient.

No documentation tool adds to your clock. Only the reporting clinician's own time counts toward the time route [3]. A scribe of any kind contributes nothing to the 30 minutes.

One thing worth stating in an article about a default. An accurate record is the point, and accuracy runs downward as often as upward. For a visit that was genuinely a 99202, a good note makes the 99202 case clearly rather than dressing it up into something else.

Sully operates across 5,000+ providers, has delivered 50M+ hours of work, and prices each role 80 to 90 percent below the human equivalent [6].

Two Worked Examples

A Visit That Is Clearly 99203

New patient, self referred after moving to the area. 36 minutes total time on the date of the encounter.

New patient confirmation: No professional service from any family medicine clinician in this group. Chart search returns no prior encounters.

Referral source: Self referred, establishing care.

Chief complaint: "I need someone here to keep managing my acid reflux."

History: 47 year old woman, gastroesophageal reflux disease diagnosed 2021, controlled on omeprazole 20mg daily for the past two years. Symptoms well controlled, breakthrough heartburn perhaps once a month. No dysphagia, no weight loss, no melena, no vomiting. No other medical history. No other medications.

Examination: Blood pressure 122/76. Heart rate 70 regular. Abdomen soft, non-tender, no epigastric tenderness, no masses. No lymphadenopathy.

Problems addressed today: Gastroesophageal reflux disease, stable.

Data reviewed: Upper endoscopy report from her previous gastroenterologist, March 2024, obtained and read. Mild erosive esophagitis at the time, no Barrett's, no dysplasia.

Risk of management selected: Continue current therapy unchanged. Low risk.

Route used: Medical decision making. Time also documented.

If time: Total time on the date of the encounter, counting only this clinician: 36 minutes, comprising review of the transferred records before the visit, the encounter itself, and documentation.

Assessment and plan: Gastroesophageal reflux disease, well controlled on omeprazole 20mg daily. Continue unchanged. Discussed trial of step-down to on-demand dosing in future, patient prefers to stay on current regimen for now. Red flag symptoms reviewed and she knows what to report. Review in 12 months, sooner if symptoms change.

One stable chronic illness puts problems at low. One external record obtained and reviewed puts data at low. Continuing an unchanged medication is low risk. That is three elements at low, which is more than the two required. The visit also clears 30 minutes, so both routes support the code, and the note states which one was used.

A Visit That Looks Like 99203 and Is Actually 99202

Same clinic, a new patient the same week. 24 minutes total time.

New patient confirmation: No professional service from any family medicine clinician in this group within three years.

Referral source: Self referred, walk-in appointment.

Chief complaint: "I've had a sore throat for three days."

History: 29 year old man, previously well, no chronic conditions, no regular medications. Sore throat three days, no fever, no cough, no difficulty swallowing, no rash. One household contact with a cold last week. No recent travel. Immunizations up to date.

Examination: Temperature 98.8 F. Blood pressure 118/72. Pharynx mildly erythematous, no exudate, no tonsillar swelling. No cervical lymphadenopathy. Chest clear.

Problems addressed today: Acute viral pharyngitis, uncomplicated and self-limited.

Data reviewed: None. Centor criteria score 0, so no rapid streptococcal test indicated.

Risk of management selected: Supportive care only. No prescription. Minimal risk.

Route used: Medical decision making.

If time: Total time on the date of the encounter, counting only this clinician: 24 minutes, comprising the encounter and documentation.

Assessment and plan: Acute viral pharyngitis. No antibiotic indicated, rationale explained to the patient including the Centor assessment. Supportive care, fluids, analgesia as needed. Return if fever develops, if swallowing becomes difficult, or if symptoms persist beyond ten days.

Why this is a 99202. One self-limited problem puts problems at straightforward, not low. No data was reviewed, and that is a clinically correct decision rather than an omission. Supportive care with no prescription is minimal risk. That is straightforward decision making.

And the time route does not rescue it. 24 minutes falls inside the 99202 band of 15 to 29 minutes, not the 99203 band of 30 to 44 [1][3].

Nothing was done wrong here. The visit was small, the clinical reasoning was sound, the note records it accurately, and the code matches. Coding it 99203 because the patient was new would be an error with no upside: it is exactly the pattern a payer's own claims data is best at finding, and it misstates a record that was already correct.

FAQ

What is CPT code 99203?

The office or other outpatient visit for a new patient requiring a medically appropriate history and examination with a low level of medical decision making, or 30 to 44 minutes of total time on the date of the encounter [1].

What are common billing errors for 99203?

The patient was not new, because someone of the same specialty in the same group saw them inside three years. The note evidences neither the time route nor the decision making route. A time figure was recorded without being qualified as total time on the date of the encounter counting only the reporting clinician. The level was chosen by default rather than read against the criteria, which produces errors in both directions. Or the submitted diagnosis does not support medical necessity.

What is the difference between CPT code 99204 and 99203?

The level of medical decision making, or the time. 99203 is low decision making, or 30 to 44 minutes. 99204 is moderate, or 45 to 59 [1]. Both are new patient visits, so the three year rule applies the same way to each.

What is the difference between CPT code 99202 and 99203?

99202 is straightforward decision making, or 15 to 29 minutes of total time. 99203 is low, or 30 to 44 [1][3]. The practical difference is usually one self-limited problem with nothing reviewed and nothing prescribed, against one stable chronic problem with something reviewed.

What is the cost of CPT code 99203?

There is no single figure. Medicare computes the fee schedule amount as the RVUs for the service multiplied by the geographic adjustment factor for your area and the conversion factor [2], so it varies by locality and by year. A non-participating supplier is paid 95 percent of the participating amount [2]. Get the current number for your locality from the CMS Physician Fee Schedule Look-Up Tool [5]. Commercial payers set their own rates separately.

Sources

[1] American Medical Association — CPT code 99203, new patient office visit, 30-44 minutes. https://www.ama-assn.org/practice-management/cpt/cpt-code-99203-new-patient-office-visit-30-44-minutes

[2] U.S. Government Publishing Office — 42 CFR 414.20, Formula for computing fee schedule amounts. https://www.govinfo.gov/content/pkg/CFR-2023-title42-vol3/xml/CFR-2023-title42-vol3-sec414-20.xml

[3] Centers for Medicare & Medicaid Services — Evaluation and Management Services, MLN006764. https://www.cms.gov/outreach-and-education/medicare-learning-network-mln/mlnproducts/downloads/eval-mgmt-serv-guide-icn006764.pdf

[4] U.S. Government Publishing Office — 42 CFR 414.22, Relative value units (RVUs). https://www.govinfo.gov/content/pkg/CFR-2023-title42-vol3/xml/CFR-2023-title42-vol3-sec414-22.xml

[5] Centers for Medicare & Medicaid Services — Physician Fee Schedule Look-Up Tool. https://www.cms.gov/medicare/physician-fee-schedule/search

[6] Sully.ai — AI Medical Scribe. https://www.sully.ai/medical-scribe

Templates

Make more room for care.

Make more room for care.