Start with the denominator. Uptake reported as one in four of all doctors billing Medicare uses as its base every clinician who submits a Medicare claim, not the smaller set practising the longitudinal care the code was written to describe [4][8]. Against the wide base, a quarter reads as brisk adoption. Against the intended base it could read as partial adoption, or as spillover, and a single share cannot separate those.
The specialty split has the same problem. It is a share of code volume, not a rate per clinician or per eligible visit, so it tells you where the dollars landed and little about whether anyone changed how they practise [5]. A specialist billing the code at a glaucoma follow-up [6] may be describing a genuine ongoing relationship, or may be adding a line item to a visit that would have happened identically in 2023. Claims cannot distinguish those, and the author does not pretend otherwise; her framing is "early signs" [14].
The arithmetic is worth having. Twenty-six million claims at $16 each is roughly $416 million billed in add-on payments in the first year [15], of which the specialist share works out near $179 million [18]. That is real money inside a fee schedule the piece describes as zero-sum [9], where rates are set with the advice of a specialty-predominant committee that has priced procedures above the visit-based cognitive work of primary care [11].
The design tension sits in the unit of payment. The uncovered work the piece identifies happens between visits: sorting ambiguous symptoms over email, or cutting a medication list back to what a patient actually needs [10]. G2211 pays only when a visit occurs, so it routes money through encounters while the deficit it names lives outside them. The nearest thing to a continuity test in the data is that in most cases the billing clinician had seen that patient more than once [7], which is a low bar, and roughly the highest bar claims data can clear.
The thing this doesn't tell you is whether care improved. There is no control arm: no matched set of visits in a world without the code, no pre-period continuity measure to difference against. What exists is first-year descriptive claims data plus interviews in which primary care physicians called the reimbursement modest and resented the extra click needed to codify work they already saw as their job [12][3]. That supports a conclusion about how the code is used, and not yet a causal one about what the code did to care. My own read, conditional on year two resembling year one: a per-visit add-on can change what a visit is called and what it costs, and it cannot buy the between-visit capacity that continuity is made of.
