Original Medicare office-visit volume shifted between common codes

Two of the most common established-patient office and outpatient visit codes moved in opposite directions in 2024 Original Medicare fee-for-service data.

Office/outpatient evaluation-and-management (E/M) codes identify billed visit levels for new or established patients. Under the CMS guidance applicable to both analyzed years, the level may be selected by medical decision making (MDM) or by qualifying total time on the date of the encounter. This analysis uses the following plain-language key:

• 99203: new patient, low MDM
• 99204: new patient, moderate MDM
• 99205: new patient, high MDM
• 99213: established patient, low MDM
• 99214: established patient, moderate MDM

Those labels describe billing categories, not a direct measure of an individual patient's severity or the duration of a particular visit. Because qualifying time is an alternative basis for code selection, the code alone does not establish how long a visit lasted.

Distinct beneficiary-day services for HCPCS 99214 rose 3.1%, from 100.6 million in 2023 to 103.7 million in 2024. The corresponding count for 99213 fell 3.1%, from 71.5 million to 69.3 million, according to MedDataUSA's analysis of national CMS rows.

The shift is visible in claims. It does not, by itself, prove that patients became sicker or visits became longer.

Quick read

• 99214 increased by 3.13 million beneficiary-day services; 99213 decreased by 2.18 million.
• Among three selected new-patient codes, 99203 fell 6.3%, while 99204 rose 1.2% and 99205 rose 2.2%.
• These counts cover Original Medicare Part B fee-for-service claims, not Medicare Advantage or the entire U.S. patient population.

Why beneficiary-day services matter

CMS publishes both line-service counts and distinct beneficiary/per-day service counts. In ordinary language, the latter counts whether a beneficiary received a particular service on a particular day within a geography, code, and place-of-service row. If the same beneficiary receives multiple units of that service on that day within the row, the measure counts one beneficiary-day service rather than every line item.

It is still not a count of unique people or globally deduplicated encounters. The same beneficiary can appear on multiple dates, under multiple codes, and in both facility and non-facility rows. We summed the national facility and non-facility rows for each code and year, so the resulting total is an aggregate of those row-level measures—not unique patients.

Reading the code shift carefully

Office and outpatient E/M codes reflect documentation and coding rules as well as clinical work. CMS guidance for 2023 and 2024 permits office/outpatient level selection based on MDM or qualifying total time. A move from one code's volume to another can arise from changes in patient mix, coding practice, site of care, program enrollment, or the number of covered encounters.

The public aggregate cannot isolate those mechanisms. It also cannot show whether the same providers changed coding patterns or whether the mix of providers and beneficiaries changed between years.

New-patient codes were mixed

The selected new-patient codes did not move as a block. Beneficiary-day services for 99203 fell from 9.23 million to 8.66 million. Code 99204 increased from 12.51 million to 12.66 million, and 99205 increased from 2.99 million to 3.05 million.

Across the five selected codes, the net increase was about 596,000 beneficiary-day services, less than one half of one percent of their combined 2023 volume. That near-flat total is consistent with redistribution among codes, but it does not prove one code directly replaced another.

The population boundary

The dataset represents 100% final-action physician and supplier Part B non-institutional line items for Original Medicare fee-for-service, subject to privacy redaction. It excludes Medicare Advantage encounters and is not a national all-payer utilization file.

For market or workforce decisions, that distinction is fundamental. A state with high Medicare Advantage enrollment can have substantial clinical activity not visible in this extract.

What to watch next

A stronger follow-up would compare the shift by geography and place of service, preserve suppressed cells, and standardize the provider and beneficiary populations where possible. Provider-level work should also avoid treating an NPI as a unique practice location.

Methodology and limitations

MedDataUSA queried the 2023 and 2024 CMS Medicare Physician & Other Practitioners by Geography and Service API versions for HCPCS 99203, 99204, 99205, 99213, and 99214. For each code and year, we required exactly one national facility row and one national non-facility row, validated all ten year/code groups as unique, then summed Tot_Bene_Day_Srvcs.

The analysis is descriptive. It does not risk-adjust, infer visit complexity, measure unique annual beneficiaries, include Medicare Advantage, or attribute causation. HCPCS descriptions are summarized for readability; official code definitions govern.

Sources

• CMS, Medicare Physician & Other Practitioners — by Geography and Service.
• CMS, Geography and Service data dictionary.
• CMS, Evaluation & Management Visits and 2024 HCPCS description changes.
• Reproducible API rows and calculations: the preserved internal analysis package.