Concinna × Agilon Health · External intelligence for value-based care

Your edge is data visibility.
We're here to extend it.

Here's the whole picture — every figure built from public federal data, no PHI and nothing to sign. Follow the path, or jump straight to whatever you want to pressure-test first.

3.56B governed records · 326 sources · 8,605 PCP-led groups · 7 yr Medicare depth · 9.5M providers · 44 named authorities · zero PHI
Scroll
How we fit

You've built your advantage on data. We're here to add to it.

Risk capture, Part D, cost trend, market selection — the levers you're already working. A public, federal, seven-year view gives each of them a second set of eyes: at physician grain, across the markets and groups beyond your own book.

It doesn't see your members — it sees the federal ground truth around them. A complement to the pipeline you've built, offered in service of the same goal.

Your screening rubric

The nine inputs you screen on — answered before the first conversation.

The standard quantitative inputs for evaluating a physician group, each produced from public federal data, seven years before you sign. Five run on the Medicare-FFS proxy used industry-wide.

Worked live for a 39-PCP independent group in one of your North Carolina markets — group de-identified; every figure reproduces against the live lake (June 2026).

Screening inputLive answer — the market & groupStatus
1MA Market Penetration61.6% of the county's Medicare is MA, 2024live
2MA Market Growth+3.7% MA enrollment, year over yearlive
3Payer Market DynamicsTop-payer share 21.5% at parent level, top-4 mappedlive
4Benchmark RateCounty $1,235.88 PMPM, CY2026live
5Group MA Membersproxy10,803 attributed FFS beneficiaries on the panellive
6Group # of PCPs39 primary-care physicians (exact, PECOS)live
7Group MA RAFproxyDrug-implied RAF 0.91live
8Group MA Revenueproxy≈$145.5M est. annual MA premium at full risklive
9Group MA MedExproxy≈$117.7M est. annual medical cost — implied margin ≈$28Mlive
The same nine — now scored and standing across 8,605 PCP-led groups in every state, not on request. Then the ownership screen runs: 7,349 are independent candidates once payer-, system-, and academic-owned groups are removed — 1,010 of them in your four core states. Pre-ranked, ownership-filtered underwriting evidence, before the first call.
The other half of the deal

Knowing the group is half the work. Knowing who owns it — and whether you can get in — is the other.

The rubric finds the right group. The corporate graph tells you whether it's independent, who controls it, and where the warm path actually runs — the read no claims feed or physician directory holds.

Worked live for a 22-PCP independent multispecialty group in a Denver-metro market — group de-identified; ownership, integrity and panel verified against the live lake (June 2026).

Access signalLive answerWhat it tells you
1Ownership typeIndependent — no public parent, not hospital-ownedAcquirable; not already inside a system
2Group shapeMultispecialty, 3 specialty lines · ~3,400-bene panelA real operating group with a referable book
3Federal integrity0 OIG / SAM exclusions across the panelClean — no integrity flags before you engage
4Corporate footprintNo holding company above it in the corporate graphNo roll-up or PE parent sitting on top (yet)
5Entity & access pathResolved on lookup — agent, standing & formation from state registriesA direct line to control — captured per group on request
And we flag who's already taken. The same lookup on comparable groups returns a national corporate agent and a “Holdings” parent — professionally managed, harder in — or a “Merged” status: already absorbed into a payer or strategic. You stop chasing the groups that are gone, and spend your access energy where it converts.
At national scale, verifiable: across 8,605 PCP-led groups we read 7,349 independent candidates and 813 already inside a payer, system, or academic parent (the rest flagged for review). The federal record confirms who is owned; confirming who is truly independent is where your own knowledge of who owns whom turns our read into proprietary truth.
Where to point it

Your core states, ranked — before you commit a market team.

Every U.S. county scored on one transparent composite — PCP supply, how little of the market is already value-based, how few REACH competitors have organized it, and risk-documentation headroom — equal-weight percentiles, no hidden weights. Cut to NY, PA, OH and NC.

MarketMedicare benesPCPsVBC densityREACH ACOsTarget score
1New York (Manhattan), NY305,6038,1930.78343.02
2Durham, NC56,1589280.8192.83
3Monroe (Rochester), NY177,7792,6860.8612.80
4Broome (Binghamton), NY47,8335470.8432.72
5Dauphin (Harrisburg), PA64,3639610.8232.62
6Buncombe (Asheville), NC68,8791,0320.83122.61
7Philadelphia, PA274,6554,9540.79182.32
The lens audits you, too. The same scoring surfaces your own Senior Health Connect ACO as the dominant REACH entity in Broome County, NY — from CMS filings alone. Computed across 3,197 counties (CMS Geographic Variation PY2024, MSSP County RAF PY2024, ACO REACH PY2023, NPPES); every figure reproduces against the live lake.
The RAF engine

An independent read on risk capture — drawn from data no single source holds.

A view of the lever you've moved most: where a panel's prescribing suggests chronic disease that coding may not yet reflect — surfaced as an opportunity to pursue, not a verdict. It comes only from triangulating four federal sources.

01 PRESCRIBINGPart D · 7 yrs 02 FDA LABELSDailyMed 03 CROSSWALKCMS-HCC V28 04 COEFFICIENTSV28 + RxHCC RAF SIGNAL RANKED
0.91
drug-implied RAF · this panel (live)
The prescribing signal flags where chronic burden is being treated but may not yet be coded — the coding-opportunity the engine ranks per panel. At ≈$900 per MA member per 0.1 RAF point, even a fraction of a point is material at panel scale.
Said plainly: this is the prescribing-implied side. We can show a physician is treating CHF, CKD and diabetes; we do not see your coding. So it is a screening signal that ranks where uncaptured risk most likely sits — the analysis we run as the engagement, not a pre-built gap list. Every drug→condition link is corroborated against its FDA label; every condition→HCC mapping is derived from CMS's own crosswalk.
What it's worth

Put your own book in.

A RAF point is worth roughly $900 per MA member, per 0.1 point, a year. Move the inputs to your book and watch the gross at stake.

Gross annual value
$383M
Gross at book scale. Realized value depends on capture — the engine ranks where it most likely sits.
The value map

Built around the levers you've prioritized.

The rubric is one use of the lake — underwriting. The same asset maps to the levers across your platform.

Risk adjustment

The last share of visibility

External, prospective RAF signal at panel grain — across existing members, ahead of reconciliation. ≈$900 per MA member per 0.1 RAF — material at book scale.

Part D

Sight before reconciliation

Seven years of Part D prescribing + RxHCC scoring + specialty-drug trajectory — an external benchmark months ahead of late-year reconciliation.

Cost trend

An external reference, earlier

The full FFS cost stack — carrier, inpatient, outpatient, per-capita by county 2014–2024 — to validate trend by market ahead of your claims lag.

Contracting

Which payer, which market

Payer share by parent × county × multi-year trend, county benchmarks, Star-bonus economics — sharper keep-or-exit decisions.

Network

Where the specialty dollars go

Every specialist by procedure, place-of-service and six-year volume; hospital-employed vs independent flagged — the margin lever, mapped.

Expansion

Selection, not volume

Pre-score every independent group in a market before the first conversation — underwriting evidence in hand, so expansion is a confident decision rather than a leap.

Quality

Defend the Star bonus

Contract-level Star Ratings and the market quality landscape — benchmark your plans and price the bonus into every decision.

Population

Chronic burden, predicted

County chronic-disease prevalence and social vulnerability fused with claims — pathway targeting, and intelligence no provider database holds.

See it live

Don't take our word for it — open a real market.

Four ways to pressure-test the lake on ground you know cold, before any conversation.

Not sure where to start? Jump back to the map ↑

How we'd start

Start small, prove it, then grow.

No leap of faith. A short pilot on your own population proves the value first — then it grows at your pace, never ahead of it.

Step 1 · Pilot

A 90-day proof, on your panel

The engine run against Agilon's own book — not a demo — on a market you name. If you continue, it credits toward the partnership. One low-risk decision, nothing more.

Step 2 · Partnership

Grow into the annual

The working engagement: the full federal-intelligence platform and the modules that matter most for risk-bearing care — risk adjustment, network, Part D, provider integrity.

Step 3 · Embedded

Wired into your systems

When the time is right, Concinna embedded directly into Agilon's own workflows — a deeper licensing arrangement, scoped to where it earns its place.

Priced to a few basis points of the medical margin it's built to protect — the full figures come with the working session, tailored to your book.

The conversation
The honest boundary. This lake will never hold your encounter data, plan bids, or member-level RAF — and it shouldn't. It's a complement to the pipeline you've built, not a replacement — an independent federal view offered alongside it, in service of the same outcomes.

"When you underwrite a new physician group — what does your profile look like, and what would seven years of their Medicare history, before you sign, be worth?"

Concinna · a 3.56-billion-row governed data lake for value-based care · powered by Talon AI LLC
100% public federal records · No PHI · Figures reproducible against the live lake