Why is your number lower than what my all-payer tool shows?

Because it answers a different question. Nevvi reports the exact count on the federal Medicare record — verifiable to the row. Not a modeled all-payer estimate. Market totals here anchor on a labeled all-Medicare estimate scaled from that observed record, with the observed count shown beside it — Medicare only, never all-payer. Modeled all-payer databases project total volume across every payer statistically; those figures typically run two to four times the observed fee-for-service foundation, with the same named providers and largely the same rank order. Same market, different denominator. How the numbers reconcile →

Can I audit a number?

Yes — every observed figure is a direct, unmodified read of CMS public files, and every page names the file and year it came from. The one estimated figure, a market's labeled all-Medicare total, is a transparent scaling of those counts you can reproduce. If a count looks wrong, check it against the public record; we encourage it. The exact sources →

What exactly do the numbers cover?

Observed procedure and hospital figures are Original Medicare (fee-for-service) claims — Part B for procedures, Part A for hospital stays — the audited floor of a market. They do not include Medicare Advantage, commercial insurance, Medicaid, or cash-pay volume. Prescribing counts are the exception: they are Medicare Part D drug-plan claims, which include Medicare Advantage plans. Market totals carry a labeled all-Medicare estimate that scales the observed foundation by each state's public FFS enrollment share — Medicare only; commercial, Medicaid, and cash-pay volume are never estimated. What the numbers mean, and their limits →

Caveats

How does Medicare Advantage change what I'm seeing?

Materially, and unevenly. Medicare Advantage procedure and hospital claims are not in the CMS public files, and MA now enrolls roughly half of Medicare nationally — so every observed procedure and hospital count here is the fee-for-service floor, not all of Medicare, and all-Medicare figures are labeled estimates scaled from it. Part D prescribing counts are the exception: they are drug-plan claims and include Medicare Advantage plans. Penetration varies widely by market, which cuts two ways: in a high-MA market the floor understates more, and raw counts from two markets with very different MA penetration are not on the same basis. That's why results pages show each market's payer mix — the FFS share and the MA trend — right next to the counts, so you can read the coverage before you read the numbers. Market totals anchor on a scaled all-of-Medicare estimate — on every tier, always labeled as an estimate and never blended into observed counts, with the observed FFS count shown beside it. The limits, precisely stated →

Are the dollar figures revenue?

No. Nevvi's headline dollar figure is what Medicare actually paid — the money that moved — and "charges" are what providers billed, not what they received. When a comparison spans states, the basis is CMS's standardized payment, which removes local fee-schedule differences. None of it is a practice's total revenue: it is the Medicare fee-for-service slice only. How to read the dollars →

Why do some cells show "—"?

Two reasons, both stated where they occur: CMS suppresses any provider×code figure based on fewer than 11 beneficiaries before publication (absence means "suppressed," never zero), and volume that cannot be unambiguously attributed — a clinician registered with several groups — is shown as "—" rather than guessed. How volume is attributed to groups →

Why doesn't a group's total match the sum of its roster?

Group totals count each group's unambiguous clinicians; clinicians affiliated with more than one group are listed in every roster they belong to — marked "member of N groups" — but their volume is never split or assigned by guesswork. Their full personal volume is on their own profile page. The attribution rule →

How current is the data?

Each year shown is CMS's most recent annual release for that calendar year; there is a natural lag between a service year and its publication. The data footer on every page names the source vintage currently in service. Sources and update cadence →

Is anything AI-generated or estimated?

No figure is AI-generated. Plain-English search uses AI for one thing only — translating your question into CPT®/HCPCS codes — and every count, charge, and ranking is then queried directly from CMS data. One figure is estimated, and it says so: a market's all-Medicare total scales the observed fee-for-service count up by each state's public FFS share — a transparent arithmetic step you can reproduce, labeled as an estimate wherever it appears, never an AI guess or a modeled projection. How plain-English search works →

What's open and what's paid?

State-by-state summaries are open to everyone. They show size, structure, and how one state compares with others. Explore opens the Overview on each code, MS-DRG, medication, provider, group, and hospital page, including charges and Medicare payments where available. Analyze adds Atlas, the Burden overlay, the Clinician network, Compare, and the What-if simulator. A Deep Dive opens the remaining detail on one group, clinician, code, MS-DRG, hospital, or medication page, including the named groups, clinicians, and hospitals shown there. It saves a dated record in Saved. Deep Dives for groups, clinicians, and codes also include Nevvi’s analysis; group Deep Dives highlight what stands out. See the plans →

Before you buy

A good first step: search a market you already know well — a group whose Medicare work you can verify — and check us against it.

Something else?

Get in touch — corrections and methodology questions are welcome.