NEVVI Medicare utilization intelligence

In 2024, South Dakota replaced knees on about six times the share of fee-for-service Medicare patients as Hawaii

Published Jul 2026 · CMS Part B PUF 📋 Cite See the data →
Knee replacements per 1,000 fee-for-service beneficiaries · by state · 2024
ME: 13MEVT: 8VTNH: 21NHWA: 18WAID: 21IDMT: 23MTND: 31NDMN: 19MNWI: 19WIIL: 18ILMI: 16MINY: 13NYMA: 15MARI: 20RIOR: 19ORNV: 16NVWY: 16WYSD: 34SDIA: 20IAIN: 24INOH: 22OHPA: 21PANJ: 15NJCT: 17CTCA: 13CAUT: 32UTCO: 26CONE: 28NEMO: 21MOKY: 16KYWV: 11WVVA: 20VAMD: 18MDDE: 24DEAZ: 21AZNM: 13NMKS: 27KSAR: 19ARTN: 22TNNC: 21NCSC: 21SCDC: 11DCOK: 19OKLA: 17LAMS: 20MSAL: 18ALGA: 19GAHI: 5HIAK: 21AKTX: 17TXFL: 19FL
Medicare knee replacements per 1,000 fee-for-service beneficiaries, by state, 2024 — counting patients, not billed lines. Full method below.
How the map moved
20192024HI: 3 → 5DC: 9 → 11WV: 10 → 11NY: 10 → 13NM: 10 → 13CA: 10 → 13ME: 12 → 13NJ: 11 → 15MA: 13 → 15NV: 11 → 16MI: 12 → 16KY: 11 → 16WY: 11 → 16LA: 14 → 17CT: 13 → 17TX: 13 → 17MD: 13 → 18AL: 15 → 18WA: 15 → 18IL: 14 → 18AR: 14 → 19OR: 14 → 19GA: 14 → 19MN: 17 → 19FL: 14 → 19WI: 16 → 19OK: 16 → 19RI: 13 → 20IA: 16 → 20VA: 13 → 20NH: 18 → 21PA: 14 → 21NC: 15 → 21ID: 21 → 21SC: 15 → 21AZ: 17 → 21MO: 15 → 21OH: 15 → 22TN: 14 → 22MT: 20 → 23IN: 16 → 24DE: 19 → 24CO: 20 → 26KS: 22 → 27NE: 23 → 28ND: 22 → 31UT: 22 → 32VT: 11 → 8MS: 10 → 20AK: 9 → 21SD: 32 → 34SD 32VT 11MS 10AK 934 SD21 AK20 MS8 VT

Between 2019 and 2024 the rate rose in 50 state markets and fell in 1, while the national rate went from 14 to 18 per 1,000 — some state markets moving far more than others.

Each line is a state: its rate per 1,000 fee-for-service beneficiaries in 2019 on the left, 2024 on the right. Rates count fee-for-service beneficiaries booked to the billing state and are not scaled to total Medicare (Medicare Advantage penetration varies by state).

South Dakota sits about 6.2 times above Hawaii here. A gap that wide is usually a question about how the figure was divided, so start there — and then be careful with what is left.

The numbers first. South Dakota records about 34 knee replacements per 1,000 people in traditional Medicare — the fee-for-service side, the part that bills Medicare directly. Hawaii records about 5, and the country as a whole about 18. The high rates cluster in the Plains.

The obvious objection is the denominator, and it is the right one. Divide by nothing and the largest states sit on top of every map, whatever their surgeons do. Each state is divided by its own fee-for-service population, which removes that — and rules out adjusting further: Medicare Advantage covers a different share of people in every state, and a per-1,000 fee-for-service rate is the one version of this figure that needs no assumption about them.

The thing being divided matters too, and here it is people. Each patient counts once on the knee-replacement claim in a year, so someone who has both knees replaced appears once. Count operations and you get a bigger number; count billed lines and bigger again. Holding that unit still in every state is what lets the spread be read side by side.

So the 6.2 times is real arithmetic. What it is not is an explanation. How old and how heavy each state's Medicare population is, how common arthritis is locally, how many orthopedic surgeons practice there, and where local thresholds for operating fall could each plausibly matter. The rate measures the variation, not its cause.

One boundary rides on the whole of it: this measures where the surgery was billed, not where the patient lives. For an operation people generally have close to home the two are near enough — but they are not the same thing, and this map is the first.

A volume record, not a verdict on need, quality or good care. A higher rate means more operations per patient in that state. It does not mean better, and it does not mean worse.

What this is

A count of how many fee-for-service Medicare patients had a knee replaced, per 1,000, by state in 2024.

What it is not

Not billed claim lines, not where patients live, not a share of all Medicare enrollees, and not all-payer.

How this is counted
Patients per 1,000 fee-for-service Medicare beneficiaries, by state, 2024 — counting patients (one per beneficiary on the total-knee-arthroplasty code), not billed lines. Books to the billing state; for a locally-delivered operation, that is where it happens. A fee-for-service rate, so not scaled to total Medicare — Medicare Advantage penetration varies by state, which is why a per-1,000 rate is assumption-free. Cells under 11 beneficiaries withheld by CMS. Source: Medicare Part B.