Policy Radar · Nº 001 · Site of Service
The Hospital Is Losing the Joint
A decade of Medicare claims, three rooms, and the quiet relocation of American joint replacement. Scroll to follow the money.
×
A decade of Medicare claims, three rooms, and the quiet relocation of American joint replacement. Scroll to follow the money.
×
Until January 2018, Medicare's inpatient-only list made the total knee a creature of the hospital ward. Every one of the half-million joints the program paid for each year came with an admission, a bed, and a hospital's inpatient price. Then CMS struck the knee from the list, and two years later the hip. What followed is one of the fastest site-of-service migrations ever recorded in the program, and it is not finished.
This is the first installment of Policy Radar, a joint Techy Surgeon × RevelAi Health series that reads CMS's own public files the way an operator has to: what changed, where it changed, how fast, and who carries the consequence. Every number on this page traces to a public dataset or a primary source, and the methodology is at the bottom, equations included.
Medicare pays for roughly half a million major joint procedures a year, and effectively all of them run through a hospital bed. The inpatient-only list is a payment rule, and payment rules decide where operations happen.
CMS removes total knee arthroplasty from the inpatient-only list. Hospital outpatient departments book 61,039 Medicare knees in year one — a setting that recorded zero the year before.
THA comes off the inpatient-only list the same month TKA joins the ASC covered-procedures list. A freestanding surgery center can now bill Medicare for a total knee. COVID empties the wards at the same time, and the elective joint never fully moves back in.
Inpatient joint discharges fall from 542,513 in 2017 to 127,759 in 2024 — down 76 percent. What remains inside the hospital is older, sicker, and increasingly fracture care rather than elective replacement.
317,419 joints in hospital outpatient departments. 70,218 in ASCs and climbing at roughly thirty percent a year. The hospital inpatient ward — the setting the entire bundle-payment apparatus was designed around — now hosts one joint in four.
Look at the totals and the story sharpens. Medicare fee-for-service volume is roughly where it was in 2019 — about 480,000 knees and hips a year, despite Medicare Advantage siphoning off enrollment all decade. Demand held. The address changed. And each address bills a different price.
ASC share of Medicare joints in 2024, state by state. The federal rule changed once, everywhere, on the same day. The map did not.
The named leaders run 21 to 27 percent of Medicare joints through freestanding centers, with Tennessee coming fastest behind them. Montana went from under 3 percent of joints to 22 in three years; the leading movers are adding three to six points of share a year.
Kansas, Alabama, Oklahoma, West Virginia: ASC share at or under 3 percent, inpatient share still above half. Certificate-of-need law, hospital employment of surgeons, and rural facility economics all act as brakes. The brake leaks: Illinois patients are 3.7× more likely than Indiana's to cross state lines for an ASC joint, taking roughly $18M a year in facility fees with them.·
Share, velocity, setting mix, payment rates, and the largest freestanding joint centers in the market. This is the radar's target layer: the markets where bundle math is being rewritten fastest.
Multiply by volume and the strategic picture writes itself. The 70,218 Medicare joints done in ASCs in 2024 would have carried on the order of $270 million more in facility payments had they stayed in HOPDs, and far more against inpatient rates. That delta is the hospital CFO's problem, the ASC investor's pitch deck, and the Medicare trustee's rare good news, all in one number. Independent claims analyses put the program's annual savings from the ASC shift in the same range.· One wrinkle belongs in the ledger, though: the patient does not always share the savings. Hospital outpatient copays are capped at the inpatient deductible while ASC coinsurance is not, so a beneficiary can owe slightly more at the dramatically cheaper facility. ASCA is lobbying to fix it.·
It is also the reason surgeons keep gaining leverage. The scarce asset in this migration is the surgeon who decides which building gets the case.
This is the question that made the inpatient-only list feel protective, and it has now been tested about as thoroughly as observational data allows. The answer keeps coming back the same way: in appropriately selected patients, same-day joint replacement is at least as safe as an overnight admission, and on several measures safer.
A propensity-matched analysis of commercial claims found readmissions, post-surgical complications, and payments all lower for outpatient joint replacement than inpatient, with ASC patients doing as well as or better than HOPD patients.· A matched comparison of same-day TKA against even the best-case "rapid recovery" overnight stay found half the minor complication rate (2.8 versus 5.8 percent), equivalent major complications and readmissions, and a $6,824 lower 90-day episode payment.· When New York's Medicare claims were examined before and after the 2018 policy change, taking the knee off the inpatient-only list was associated with no deterioration in readmissions or emergency visits at all.· High-volume centers running mature same-day pathways report 97 percent same-day discharge success with readmission rates under one percent.· Patient-reported outcomes track the same: hip and knee scores are indistinguishable between same-day and admitted cohorts.·
Two honest caveats belong next to that conclusion. First, selection is doing real work: the patients going home the same day are chosen because they can, and the sicker joint stays admitted, which is exactly why the remaining inpatient mix keeps getting more expensive. Second, the choosing is not colorblind. Older, Black, female, and safety-net patients are measurably less likely to be offered the outpatient pathway,· and the freestanding centers absorbing this volume treat a whiter, healthier, more privately insured population while hospital-owned settings keep the complex and the vulnerable.· Meanwhile the risk profile of outpatient joint patients keeps drifting upward each year as programs stretch their criteria.· Who gets the $6,900 operation with the better recovery, and who gets the $16,000 admission, is becoming its own equity question. The answer is to widen the doorway rather than narrow it: criteria-based selection instead of payer-mix selection, preoperative optimization that makes more patients same-day eligible, and caregiver, transport, and home-support pathways for those screened out by circumstance rather than physiology. Policy carries a matching obligation: hospitals still holding the complex book, many of them safety nets, should be funded and given a runway to build same-day programs, not judged against migration-adjusted benchmarks they lack the capital to chase.
TEAM Went Live In January 2026. Site-Of-Service Mix Is Now The First Thing To Model In Every Episode.
The policy machinery that will decide how far the migration runs. Tracked against the primary documents.
Built from CMS public use files via the data.cms.gov API. Reproducible, falsifiable, and honest about its edges.
If you run episodes under TEAM, the site-of-service mix above is already moving your target prices. RevelAi models it, then builds the care coordination that captures it.