A Medicare Code Meant to Pay for Primary Care Relationships Paid Out $394 Million. Specialists Billed 43 Percent of the Claims.

Analysis by DPC Insider — Research Desk
#medicare#primary-care-payment#fee-for-service#research#policy
Sources: A billing code was supposed to incentivize relationship-based primary care for Medicare patients. It may not be working as hoped · Billing of Medicare's G2211 Longitudinal Care Code Among Traditional Medicare Beneficiaries · Early Use of Medicare's G2211 Code Highlights Challenges in Supporting Longitudinal Care · Docs Land Nearly $400 Million in First Year of Controversial Medicare Billing Code · Advanced Primary Care Management · Advanced Primary Care Management (AAFP) · Direct Primary Care (AAFP)

A Medicare billing code built to reward the continuing relationship between a patient and the doctor who knows them paid out roughly $394 million in 2024. Specialists billed 43 percent of the claims.

That accounting comes from a JAMA research letter published February 19, 2026 by Ishani Ganguli and colleagues, a national accounting of how Medicare’s G2211 code behaved once it was loose in the world. Ganguli, a primary care physician and health services researcher at Brigham and Women’s Hospital and Harvard Medical School, returned to the finding on August 28 in a STAT opinion piece with a blunter headline. The code may not be working as hoped.

What G2211 Was Built To Pay For

CMS switched on G2211 on January 1, 2024. It is an add-on rather than a standalone service. A clinician appends it to an office visit claim to mark that the visit sat inside ongoing, relationship-based care instead of a one-off encounter. Ganguli puts the payment at about $16 per visit, with roughly $3 of that landing on the patient as cost sharing.

The theory was clean. Fee-for-service pays for discrete work, and almost nothing that makes primary care valuable is discrete. Adding a small premium to every visit inside a continuing relationship was meant to tip the money toward the doctors carrying those relationships.

Adoption was not the problem. In its first year the code was billed 26 million times for 10.6 million traditional Medicare beneficiaries, and Ganguli reports that about one in four of all doctors billing Medicare used it. The STAT piece calls that good uptake next to prior similar codes.

The Split

The distribution is where the design shows.

Of those 26 million codes, specialist physicians billed 43.0 percent. Primary care physicians billed 39.7 percent. Other clinicians accounted for the remaining 17.4 percent.

Taken alone, that reads like specialists helping themselves to an add-on written for somebody else. The rest of the table complicates it. Primary care physicians attached G2211 to 24.1 percent of their eligible visits. Specialists attached it to 13.1 percent, and other clinicians to 8.9 percent. Per eligible visit, primary care used the code close to twice as often. Specialists still billed it more times overall. What the letter does not print is how many eligible visits each group had. It does print enough to estimate one: a group’s share of the 26 million codes set against its rate per eligible visit implies its pool, and on the letter’s own figures the specialists’ pool comes out roughly twice the size of the primary care pool. That points at the larger denominator rather than heavier use, though the letter stops short of saying so itself.

The billing went where the visit volume is. A code that pays per visit will do that every time, whoever the drafters had in mind.

Ganguli’s STAT piece supplies the qualitative version of the same problem, describing specialists billing the code for acid reflux and mild glaucoma, conditions that sit awkwardly against what the code says it is for. The research letter also flags that it could not identify specialty for nonphysician clinicians, so that 17.4 percent bucket stays partly opaque.

Thirty-Seven Dollars a Year

Divide the payments by the patients and G2211 delivered about $37 per beneficiary across all of 2024. That is the whole year’s payment, per patient, for keeping primary care continuous.

Set it beside what a Direct Primary Care patient pays out of pocket. The AAFP says monthly membership fees for DPC patients generally range from $50 to $100. Whatever a reader makes of that price, the two figures are not measuring the same category of thing.

The gap is structural rather than a matter of Medicare being stingy. G2211 requires a visit to exist at all. A membership is paid whether or not the patient walks in, which is what buys the text answered at 9pm, the medication adjusted by phone and the forty-minute visit that fee-for-service would price at fifteen.

Medicare Already Built the Other Version

CMS did not wait for the JAMA numbers. On January 1, 2025 it activated three Advanced Primary Care Management codes, G0556, G0557 and G0558, tiered by chronic condition count and by Qualified Medicare Beneficiary status. They pay monthly, per patient. An initiating visit is required for patients who are new to the practice, have not been seen there within the previous three years, or have not had another care management service from the practice within the previous year. The AAFP notes that CMS eliminated the time thresholds the earlier care management codes ran on. The practice also has to be able to deliver care outside the traditional office visit, through phone, e-visits, home visits or extended hours.

That is a per-member-per-month payment for a defined panel, run by Medicare. Ganguli writes that CMS plans to explore a more ambitious version of these payments in 2027.

The Counterargument

The obvious objection to reading G2211 as a failure is that it plainly wasn’t one. 26 million billings in year one is uptake the STAT piece itself calls good next to prior similar codes. Primary care physicians used it more intensively than any other group per eligible visit. They billed close to 40 percent of the codes. Calling that a flop because specialists also qualified is grading against a target CMS never set.

The objection holds on adoption and loses on design. G2211 was supposed to make continuity pay. What it paid for was visits occurring inside continuity, a narrower thing, and volume then went where volume lives. The strongest evidence for that reading sits outside the JAMA table entirely. Medicare shipped the monthly version twelve months later, and it is now sizing up a bigger one for 2027. Agencies do not replace instruments that are working.