CMS ACCESS accepted-applicants roster — re-attempt to close the unnamed-~100 blind spot on the 3-leg (DS+VBC+PI) rubric
The question
Pull the full CMS ACCESS Model accepted-applicants roster (HTML or PDF from cms.gov; WebFetch timed out twice previously) and check the unnamed ~100 against the (DS + VBC + PI) 3-leg rubric for any 3/3 patient-data-sovereignty competitor. Context: the parent [[2026-05-18-cms-access-approved-applicant-roster]] named and cleared ~50 of the 150+ applicants; the residual risk lives in the unnamed ~100.
What we already know (from the vault)
- [[2026-05-18-cms-access-approved-applicant-roster]] (parent) — verdict no 3/3 competitor in the ACCESS 150+ roster; every named participant scored at most 1/3 (VBC leg present, DS and PI absent). It named ~50 across trade press and explicitly left the unnamed ~100 as the residual risk, seeding this exact re-attempt ("the WebFetch timed out twice... try a direct curl or the Wayback snapshot").
- [[2026-05-11-patient-data-sovereignty-competitor-scan]] — origin of the 3-leg rubric. DS = patient-as-data-source-of-truth / data-principal; VBC = value-based contract with a payer; PI = patient cash payouts from the savings share. A 3/3 = a direct competitor to RDCO's patient-data-sovereignty thesis. 12 candidates scored, none 3/3 ("clean white space").
- [[2026-07-06-lead-eom-py2026-patient-data-sovereignty-competitor-rubric]] — applied the identical rubric to the adjacent CMS venues (LEAD, EOM); no 3/3, and documented why it's structural: these models pay the provider/ACO for savings, not the patient. Same 1/3 ceiling as ACCESS.
- [[2026-06-27-patient-data-sovereignty-competitor-pivot-v2]] and [[2026-07-09-patient-data-rights-startups-2024-2026-scan]] — the three closest analogs have pivoted away from the wedge; the field is thinning, not filling. Context for reading "no 3/3" as durable, not a snapshot.
- [[2026-07-06-aledade-partner-terms-access-model]] — Aledade (in ACCESS) scored 1/3, a partner shape not a competitor shape.
What the web says
- The canonical CMS roster stayed unreachable to automated fetch — again. cms.gov/priorities/innovation/access-model-accepted-applicants returned HTTP 403 Forbidden on this attempt (prior attempts timed out; the LEAD/EOM brief hit the same 403 wall on cms.gov). The page was last updated May 27, 2026 per search-index metadata (roster grew on rolling review from the April ~150).
- No Wayback snapshot exists. The Internet Archive availability API returns empty
archived_snapshotsfor the accepted-applicants URL — no cached copy to scrape. - No machine-readable dataset found. A targeted search for a data.cms.gov "Innovation Center Model Participants" CSV for ACCESS surfaced nothing; ACCESS only launched July 5, 2026, so a participant dataset may not be published yet.
- The trade-press enumerations reproduce the same ~50 subset, not the full 150+. Longyear Health's analysis is the fullest enumeration and explicitly calls itself "a first look... a subset." It adds legal-entity granularity (Aledade Care Solutions of California PC / Delaware LLC; Cadence Health CA/FL/NJ entities; Pair Team CA/PA entities) but zero new company-level names beyond the parent brief's ~50.
- One previously-unnamed applicant surfaced via Second Opinion: Story Health Partners (virtual cardiovascular / heart-failure specialty care). Also confirmed explicitly NOT participating: Hinge Health, Sword Health, Oura, Apple. The article's "full list" is paywalled; it displays only a subset.
- Structural confirmation of the data-flow direction. In the Second Opinion coverage, Lisa Bari (Innovaccer) frames ACCESS incentives around cost reduction and data-sharing back to health systems — provider-ward, not patient-ward. No named org in any surveyed source positions on patient data ownership, patient-as-data-principal, health-data sovereignty/consent, or patient cash-from-savings.
- The ASC News MSK-track article 403'd; the MSK names it would carry (Bold, Plethy, Flagler, JOGO Health, Livara, RightMove, Redefine) are already in the cleared ~50.
Convergences and contradictions
- Convergence: four independent enumerations (Longyear, Second Opinion, plus the Healthcare Dive / Fierce / BHB set from the parent brief) name overlapping subsets and surface no sovereignty-positioned company. The one new name (Story Health) is a VBC specialty-care delivery play, not a sovereignty play.
- Convergence (structural): the Innovaccer framing (data flows back to health systems) matches the LEAD/EOM finding — CMS value-based models reward the provider/ACO for savings, so a participant can max the VBC leg and stay permanently at 1/3. The white space is structural, not incidental.
- Contradiction with the task's premise: the roster was assumed pullable "HTML or PDF from cms.gov." In practice cms.gov hard-blocks automated fetch (403), there is no Wayback copy, and the only canonical full enumeration (STAT+ / the CMS page itself) is paywalled or blocked. The unnamed ~100 could not be individually enumerated this pass either.
Synthesis for RDCO
No unnamed applicant was found to score 2/3 or 3/3 — but that is because none of the unnamed ~100 could be individually enumerated, not because they were each read and cleared. This pass did not obtain the roster. The direct CMS page returned a hard 403 (the "timed out twice" of the parent brief is now "403-blocked" — same net result: unreachable to automated fetch), the Internet Archive has no snapshot, and no machine-readable data.cms.gov dataset for ACCESS exists yet. I did not fabricate any of the ~100 names, and none should be inferred.
What did move: the named-and-cleared set expanded by one. Story Health Partners is newly surfaced and scores 1/3 — VBC leg present (at-risk virtual cardiovascular specialty care, the exact ACCESS shape), DS absent (standard EHR data, not patient-as-data-principal), PI absent (no patient cash from a savings share). It is the same 1/3-ceiling pattern as every other named participant, and it is a partner/adjacent shape, not a competitor shape. Second Opinion also confirmed that the marquee wearable/MSK names one might most expect (Hinge, Sword, Oura, Apple) are not in ACCESS — narrowing, not widening, the field of plausible sovereignty-adjacent entrants.
So the blind spot is narrowed, not closed. The narrowing rests on three legs of indirect evidence rather than a roster read: (1) four independent enumerations now cover ~51 of the 150+ and surface zero sovereignty-positioned names — the sampling keeps returning the same clusters (consumer wellness, behavioral health, VBC enablers, food-as-medicine, AI-native chronic care, MSK, FQHCs/health systems); (2) the structural argument is strong and now doubly confirmed — ACCESS pays the provider for savings and routes data back to health systems, so a participant would have to bolt DS+PI on from outside the model, and nothing observed does that; (3) the June 27 close-out shows the closest competitors pivoting away, so the prior probability of a hidden 3/3 is falling over time, not rising. The honest residual risk: ~100 applicants remain individually unread, and a sovereignty-positioned org small enough to escape all trade-press coverage could sit among them. That probability is low and getting lower, but it is not zero, and this pass did not drive it to zero.
Net for the bet: the patient-data-sovereignty + outcome-procurement wedge survives this re-check. ACCESS remains a venue to operationalize the wedge inside, not a crowded field. But the "pull the full roster" task should be marked still open / blocked on access, not complete — the correct next action is a non-WebFetch retrieval path (see follow-ups), not another WebFetch retry, which will keep 403-ing.
Why this is in the vault
Keeps the patient-data-sovereignty + outcome-procurement wedge honest: it records that the unnamed ~100 CMS ACCESS applicants remain individually unread (cms.gov now 403-blocks automated fetch), so the competitor blind spot is narrowed but NOT closed. Without this note the roster task would be mis-marked complete and the residual risk would drop silently off the bet's risk register.
Open follow-ups
- Retrieve the CMS roster via a non-WebFetch path that survives the 403: a browser-context pull (Playwright MCP with a real user-agent) of cms.gov/priorities/innovation/access-model-accepted-applicants, or a
curlwith browser headers. This is the single action most likely to actually close the blind spot. - Check data.cms.gov for an "Innovation Center Model Participants" / ACCESS participant dataset now that the model launched July 5, 2026 — a CSV would enumerate all 150+ at once.
- Single-entity scoring pass on Story Health Partners if it ever adds a patient-incentive or patient-data-principal layer (currently 1/3; watch for 1/3→2/3 velocity, same flag as PicnicHealth ThumbPrint).
- Re-run against the Jan 1, 2027 ACCESS cohort (post-May-15-2026 applications) — the wedge is most exposed if a 2026 sovereignty-positioned entrant applies in the second window.
- Trigger the LEAD Model roster check ([[2026-07-06-lead-eom-py2026-patient-data-sovereignty-competitor-rubric]]) once CMS publishes the LEAD accepted-applicant list ahead of the Jan 1 2027 launch — still deferred, no roster exists yet.
Related
- [[2026-05-18-cms-access-approved-applicant-roster]]
- [[2026-05-11-patient-data-sovereignty-competitor-scan]]
- [[2026-07-06-lead-eom-py2026-patient-data-sovereignty-competitor-rubric]]
- [[2026-06-27-patient-data-sovereignty-competitor-pivot-v2]]
- [[2026-07-09-patient-data-rights-startups-2024-2026-scan]]
- [[2026-07-06-aledade-partner-terms-access-model]]
- [[2026-05-10-data-sovereignty-outcome-procurement-bet-architecture]]
Sources
- [[2026-05-18-cms-access-approved-applicant-roster]] — parent brief, rubric + named ~50 (
~/rdco-vault/06-reference/research/2026-05-18-cms-access-approved-applicant-roster.md) - [[2026-05-11-patient-data-sovereignty-competitor-scan]] — DS+VBC+PI rubric origin (
~/rdco-vault/06-reference/research/2026-05-11-patient-data-sovereignty-competitor-scan.md) - [[2026-07-06-lead-eom-py2026-patient-data-sovereignty-competitor-rubric]] — adjacent-venue check, structural argument (
~/rdco-vault/06-reference/research/2026-07-06-lead-eom-py2026-patient-data-sovereignty-competitor-rubric.md) - [[2026-06-27-patient-data-sovereignty-competitor-pivot-v2]] — competitors pivoting away (
~/rdco-vault/06-reference/research/2026-06-27-patient-data-sovereignty-competitor-pivot-v2.md) - [[2026-07-09-patient-data-rights-startups-2024-2026-scan]] — recent residual blind-spot sweep (
~/rdco-vault/06-reference/research/2026-07-09-patient-data-rights-startups-2024-2026-scan.md) - [[2026-07-06-aledade-partner-terms-access-model]] — Aledade partner-vs-threat (
~/rdco-vault/06-reference/research/2026-07-06-aledade-partner-terms-access-model.md) - [[2026-05-10-data-sovereignty-outcome-procurement-bet-architecture]] — bet architecture (
~/rdco-vault/01-projects/health-and-longevity/2026-05-10-data-sovereignty-outcome-procurement-bet-architecture.md) - CMS ACCESS Model Accepted Applicants (canonical) — HTTP 403 to automated fetch this pass; last updated May 27, 2026 per search index
- Longyear Health — Analyzing the CMS ACCESS Model Participants List — fullest enumeration, explicitly a subset; no new company-level names
- Second Opinion — CMS announces ACCESS participants — surfaced Story Health Partners; confirmed Hinge/Sword/Oura/Apple NOT participating; Innovaccer/Lisa Bari data-flow framing
- ASC News — Who made the Musculoskeletal Track — HTTP 403 to automated fetch
- Internet Archive availability API — no archived snapshot of the roster page