July 12, 2026
There is no national data on APCM uptake. So we measured the next best thing.
CMS has published no utilization data on its newest care-management program. Our analysis of Medicare public use files shows what the activation gap actually looks like — state by state.
Share with others
Table of contents
Why there’s a data hole
Advanced Primary Care Management (G0556, G0557, G0558) went live January 1, 2025 — a flat monthly per-patient payment for between-visit chronic care work, tiered by complexity instead of tracked by minutes (roughly $16, $54, and $117 per patient per month at 2026 national averages; the full code and rate breakdown lives in our APCM guide). It is the biggest structural change to Medicare care-management billing in a decade.
And nobody knows who’s using it. No CMS utilization release, no MedPAC analysis, no peer-reviewed estimate exists as of June 2026. CMS itself used the CY2026 proposed rule to issue a Request for Information on how APCM is being used — the agency asking the public, rather than reporting. On the RHC/FQHC side, NARHC’s published position is only that “anecdotally, some RHCs have begun billing for APCM codes.”
So we measured the next best thing: the predecessor behavior. Using the CMS Medicare Physician & Other Practitioners public use files, we computed how many primary-care physicians billed even one Chronic Care Management code (the 99490 family) in CY2024 — the last full year before APCM, released by CMS in May 2026. Same patients, same between-visit work, same consent requirement.
What we found
- 5.1% of US primary-care physicians (9,173 of 178,495 in Family Practice, Internal Medicine, and Geriatrics) billed any CCM code in CY2024.
- Across the 14-state rural belt: 6.4%. Among rural-belt physicians practicing in rural (RUCA 4+) communities: 7.3%.
- Put differently: in rural America, roughly 14 primary-care physicians are not billing care management for every 1 who is (13.8:1 nationally; 12.7:1 inside the belt).
- The five-year trend is glacial: 4.3% (CY2019) to 5.1% (CY2024). A fee-schedule code left to activate on its own moves about 0.16 percentage points per year.
This is consistent with the published beneficiary-side benchmark — CCM reached only 3.4% of eligible beneficiaries by 2019 (Jang et al., Journal of the American Geriatrics Society, 2024). And it matches what we hear in the field: among rural practices that already know about APCM specifically, we see roughly a 4:1 gap between aware and billing. That last figure is a CareAtlas field observation, not a published statistic — but it makes the point: awareness shrinks the gap and does not close it.
Activation is wildly uneven — the state table
Physician CCM activation, CY2024 (CareAtlas analysis; rural = provider address RUCA 4–10):
Two things stand out. First, rural physicians out-activate metro — 6.7% versus 4.9% nationally. Rural primary care is already leading adoption, and is still more than 93% unactivated. Second, the spread — Iowa tripled to 15.5% while North Dakota sits at 0.8%, on an identical fee schedule — says activation is an operations phenomenon, not a policy one. The code pays the same everywhere. What differs is whether anyone stands up the program.
Why the gap is workflow, not awareness
A decade of CCM established that awareness was never the constraint. The published record and our field experience converge on three activation barriers:
- The consent and cost-sharing conversation. APCM requires documented patient consent, and the ~20% coinsurance is the most consistently cited enrollment barrier in published commentary. The notable exception: QMB (dual-eligible) patients cannot be billed cost-sharing — so the panels with the highest-rate tier also sidestep the hardest barrier.
- Program requirements without program staffing. APCM’s 13 service elements assume risk stratification, 24/7 access arrangements, and population-level management. A 3-provider practice hears “care model redesign,” not “new code.”
- Billing-system transition drag. Post-G0511, RHCs and FQHCs had to pick one billing method and apply it consistently; MAC-level confusion through 2025 made some shops wait it out. (What replaced G0511, in full.)
What separates the practices that bill from the ones that don’t, in our observation: an existing CCM or RPM workflow, a panel with meaningful QMB share, and one named owner for consent capture. None of these are awareness problems — which is why they’re solvable without adding staff.
Methods, in one paragraph
Numerator: unique individual NPIs billing any of 99490/99439/99487/99489/99491, from the CMS by Provider and Service PUF. Denominator: individual clinicians in Family Practice, Internal Medicine, or Geriatrics with any Medicare Part B claims that year, from the by Provider PUF (nurse practitioners computed separately: 2.2% national). Rural = provider-address RUCA 4–10. Three caveats we’d want a referee to know: the PUF suppresses provider-code rows with 10 or fewer beneficiaries, so every rate here is a conservative undercount; RHC/FQHC bundled billing (G0511) moved on institutional claims and is not captured, so these rates describe carrier-billing clinicians only; and CCM is a proxy baseline for APCM — same patients, same between-visit work, same consent requirement, harder documentation. Full methods, the 50-state table, and the underlying extracts are available on request.
