Open Health Data Hub



All posts
|3 min read

Knee and Hip Replacement Costs Vary by 80% Across States

joint replacement Medicare costDRG 470 spendinghip knee replacement price variationMedicare inpatient geographic variationelective surgery cost disparity

Joint replacement is supposed to be Medicare's showcase for value-based care. A decade of bundled payment experiments, quality reporting, and cost benchmarking has targeted exactly this procedure. Yet in 2023, Maryland Medicare pays $23,530 per knee or hip replacement discharge while the national weighted average sits at $13,695. That's not a rounding error. It's a 2.49-to-1 ratio between the most and least expensive states.

Maryland and Alaska Sit in a Different Payment Universe

The state-level spread for DRG 470 is stark. Maryland's 29 hospitals averaged $23,530 per discharge in 2023. Alaska came second at $19,531. California, with 130 hospitals and far more volume, averaged $17,525. Minnesota, ranked 20th, averaged $13,367, less than 57 cents on the dollar compared to Maryland.

The total payment ratio (Medicare plus beneficiary cost combined) is somewhat tighter at 2.09-to-1, which means the Medicare payment gap is wider than the total cost gap. That compression matters: it implies beneficiary out-of-pocket costs are partly absorbing the difference, not just program spending.

For Medicare's budget, the geographic spread translates directly into billions of dollars in differential spending for a procedure performed at high volume across the country. A beneficiary in a high-payment state isn't necessarily getting a better implant or a better outcome.

Ten Years of Reform, and the Spread Is Wider

Here's what makes the state-level variation harder to dismiss: it's getting worse, not better.

The coefficient of variation for hospital-level Medicare payments was 25.1% in 2013. By 2023, it had risen to 28.0%. The standard deviation of hospital-level payments grew from $3,143 to $3,784 over the same period. The national weighted average payment rose from $12,088 in 2013 to $13,695 in 2023, a 13.3% increase in nominal terms. Prices went up and spread out simultaneously.

This is the decade that produced the Bundled Payments for Care Improvement program, the Comprehensive Care for Joint Replacement model, and sustained pressure on post-acute spending. None of it produced convergence. The variation that existed when the Affordable Care Act's delivery reform provisions were just getting started is still there, slightly larger, in 2023.

Patients in New Mexico and Alabama Carry the Heaviest Out-of-Pocket Load

The Medicare payment story is about program spending. The out-of-pocket story is about patients.

New Mexico had the highest average beneficiary out-of-pocket cost per discharge at $4,900, representing 29.6% of the total payment of $16,568. Alabama's burden was even more striking in proportional terms: beneficiaries there paid an average of $4,551 per discharge, representing 32.0% of total payment across 1,188 discharges. Compare that to California, where out-of-pocket costs averaged 14.9% of total payment. A New Mexico Medicare beneficiary getting a knee replacement pays roughly twice the share of total cost as a California beneficiary getting the same procedure.

The geographic pattern of out-of-pocket burden doesn't simply mirror the geographic pattern of total costs. New Mexico's total payment of $16,568 is well below Maryland's $26,111, but New Mexico patients pay a far higher share. That means low-cost states aren't necessarily easier on patients' wallets.

For a Medicare beneficiary on a fixed income, a $4,900 out-of-pocket bill for a single procedure is a significant financial event. The variation in that number across state lines, for the same DRG, is a direct measure of how unevenly the program's cost-sharing structure lands on patients depending on where they happen to live.

Given that the coefficient of variation in hospital-level payments held at 25.1% in 2013 and remained at 28.0% a full decade later, the question that the data raises but doesn't answer is what, exactly, would have to change for price convergence to actually happen.

Explore the data yourself

Run your own queries against 240M+ rows of federal health data using natural language — powered by AI.

Start analyzing