06-reference/research

cms access model oap calibration

2026-09-02·research-brief·source: deep-research·by Ray Data Co (deep-research synthesis)
healthcarecms-access-modeloutcome-aligned-paymentvalue-based-carepatient-data-sovereigntyrate-setting

CMS ACCESS Model OAP calibration — the rate card is published, and it is a fee schedule, not a savings pool

The question

"CMS ACCESS Model OAP calibration. Parent brief flagged this as 'Federal Register watch.' Now that 150+ applicants signed up, OAP rate-setting is no longer hypothetical. Action: schedule a follow-up scan once the first cohort starts paying (Aug-Sep 2026)."

Follow-up #6 from [[2026-05-11-patient-data-sovereignty-competitor-scan]], scheduled for exactly this window. The bet architecture behind it prices a patient-data-sovereignty care platform against ACCESS Outcome-Aligned Payment (OAP) economics, so the actual dollar calibration decides whether the three-leg wedge is financeable inside ACCESS.

What we already know (from the vault)

What the web says

Convergences and contradictions

Synthesis for RDCO

The headline correction: ACCESS cannot fund the patient-payout (PI) leg of the wedge, and the bet architecture should stop assuming it can. [[2026-05-19-aledade-patient-incentive-integration]] listed "the ACCESS Model OAP framework (which has more flexibility)" as escape hatch (b) from the $20 Beneficiary Inducement CMP cap. The published rate card closes that hatch. A CKM Initial-Period beneficiary generates a maximum of $336 in Medicare payment for a full year ($420 allowed × 80%), half of it withheld until post-period reconciliation, dropping to $168/year in the Follow-On Period. There is no avoided-spend upside; the only spend-linked mechanism, the Substitute Spend Adjustment, is a penalty. A three-way savings split has nothing to split. Whatever funds the patient cash leg, it will be commercial/MA contracts or an Innovation Center waiver — escape hatches (a) and (c) — never ACCESS OAP.

The one patient-side lever ACCESS does sanction is the coinsurance waiver, and it is small but clean. Footnote 1 permits participants to adopt a uniform policy waiving the beneficiary's 20% coinsurance. That is a model-sanctioned, uniform-policy patient-side economic benefit worth up to $84/year (CKM Initial), $72 (eCKM Initial), $36 (MSK or BH Initial). It is in-kind and capped, not a cash outcome share, and it is a cost to the participant rather than a distribution from a pool — but it is the only patient-side transfer inside the model that is pre-cleared rather than requiring an AKS analysis. It should be modelled as a CAC offset, not as the PI leg: against the $50-200 payer-attributed acquisition range in [[2026-07-06-patient-side-cac-data-sovereignty-care-brand]], a waived $84 is a material enrollment inducement at roughly half the cheap-mode CAC, which likely explains why CMS pre-authorized it.

Re-sizing the bet. A 100K-beneficiary CKM panel maxes at ~$33.6M of Medicare revenue in a pure-Initial-Period year and ~$16.8M in a pure Follow-On year, before any Clinical Outcome or Substitute Spend haircut and before the 5% multi-track discount. The $25-50M figure in the origin brief survives numerically for year one and then halves — but it now represents the ceiling of a fee schedule rather than the midpoint of a savings share, which changes everything about the risk profile: no upside tail, downside only through adjustments, and the panel must be continuously replenished with Initial-Period patients to hold revenue flat. Combined with [[2026-07-06-aledade-partner-terms-access-model]]'s finding that ACCESS carries no capital-reserve mandate and no downside payment adjustment, ACCESS reads as a low-risk, low-ceiling, volume-and-throughput business — attractive as a proving ground, structurally incapable of being the wedge's economic engine.

The genuinely new strategic opening is the multi-payer alignment kit, and it points at Variant A. CMS is publishing track-specific G-codes usable by any payer, a payment-adjustment reference implementation, a provider-agreement template, and a FHIR-based OAP Measure reporting API — with private payers covering 165M lives committed to aligning. That is CMS deliberately turning OAP measurement into a cross-payer standard. The durable, reusable asset in this whole venue is therefore not a care panel; it is the OAP measurement and attestation pipeline: PROM capture (PGIC, WHODAS 2.0, NRS, PROMIS PF/PI, QuickDASH, NDI, ODI, HOOS JR, KOOS JR), biomarker ingestion against per-beneficiary baseline-relative improvement targets, on-time end-of-period submission (a missed measure is scored as non-attainment, so reporting reliability is directly revenue-bearing), Outcome Attainment Rate computation against the 50% OAT, and Substitute Spend exposure monitoring. 205 participants — most of whom, per CMS, have never served Medicare beneficiaries — all need that pipeline, all face the same 50% withhold, and all will face the same requirement again from aligned commercial payers. This is a much more concrete Variant A (platform/SaaS) product thesis than "sell the missing patient-side leg to Aledade-shape operators," it has a named buyer universe of 205 organizations with a published deadline structure, and the patient-data-sovereignty layer can be built underneath it later rather than sold up front.

Why this is in the vault

It replaces the two load-bearing ACCESS assumptions in [[2026-05-10-data-sovereignty-outcome-procurement-bet-architecture]] and its parent thesis — the "~$100/beneficiary/year OAP floor" and the "Global-style 50% shared savings" revenue model — with the published CMS rate card, and it closes the "map the ACCESS OAP rate card for CKM" follow-up left open by [[2026-07-06-aledade-partner-terms-access-model]]. It also retires the "Federal Register watch" trigger that three briefs in this chain are still waiting on, and corrects a factual error (OAT escalation) that would misprice any multi-year ACCESS model.

Open follow-ups

  1. Where does the patient cash leg actually get funded? With ACCESS ruled out, price escape hatches (a) commercial/MA contracts outside BIP rules and (c) an Innovation Center §1115A waiver. Which one has a live precedent for outcome-tied cash to beneficiaries at above-$20 magnitude?
  2. Confirm the referring-clinician co-management payment. The CMS PCP-facing ACCESS page could not be located (two URL guesses 404'd); the "$30 per service" figure is unverified. Pin the code and amount from a primary CMS source — it is a second revenue line for any RDCO-adjacent entity that also holds referring relationships.
  3. Watch for payment-amounts v11. The current document covers only through 2026-12-31 → 2027-12-31. Track the version number and "Model Updates" block on the cms.gov ACCESS page; a v11 with different track amounts, a changed OAT, or a changed SST would re-price everything above.
  4. Which of the 165M-member aligned private payers have published their ACCESS-aligned rates? Commercial/MA OAP-aligned rates are the venue where a patient savings-share is legally possible. Named payers plus their rate cards would size escape hatch (a) directly.
  5. First reconciliation is ~July 2027, not 2026. The first 12-month Care Periods from the July 5, 2026 cohort conclude mid-2027, so the first Clinical Outcome and Substitute Spend adjustments — and the first real evidence of what fraction of participants clear the 50% OAT — land then. Schedule the observability scan for Q3 2027, not sooner.
  6. Does the FHIR-based OAP Measure reporting API have public specs? If CMS publishes the API contract and the payment-adjustment reference code, the measurement-pipeline product can be scoped against a real interface rather than inferred. Check the ACCESS Technical FAQs and any CMS developer/implementation-guide surface.
  7. Retire the Federal Register trigger across the chain. [[2026-05-10-healthcare-outcome-procurement-pioneering-provider]] and any derivative still watching the Federal Register for CMMI model parameters should be amended to watch cms.gov sub-regulatory guidance instead — CMMI model rate-setting does not appear to route through notice-and-comment at all.

Methodological note worth keeping: cms.gov returns HTTP 403 to WebFetch (and did so to the May 18 and July 13 attempts), but serves the same pages and PDFs to a plain curl -sSL -A "Mozilla/5.0". The accepted-applicants roster and the 21-page payment-amounts PDF were both retrieved that way. Use curl-plus-user-agent for cms.gov from here on.

Related

Sources

Vault:

Web (accessed 2026-09-02):