Same DRG, wildly different price: hospital charge variation, 2024

A DRG code fixes two things and leaves a third loose. Medicare assigns each inpatient stay one MS-DRG — a Medicare Severity Diagnosis-Related Group — and pays a single administratively set rate for it. The code is fixed; the payment is fixed. What a hospital charges for that same code is not. Each hospital sets its own list price on its own chargemaster, and those prices diverge far more than the clinical sameness of the code would suggest.

This study measures that divergence. It aggregates the CMS Medicare Inpatient Hospitals by Provider and Service public-use file — one row per hospital per DRG — and, for every high-volume MS-DRG, computes the spread of billed charges across hospitals: the 10th percentile, the median, the 90th percentile, and the ratio between them. It is the hospital-to-hospital companion to the DRG cost reference, which ranks what each code pays nationally. Here the question is narrower and sharper: for one code, how far apart do hospitals price it?

How wide is the spread

Across 219 high-volume MS-DRGs in 2024, the 90th-percentile hospital billed a median of 3.8× what the 10th-percentile hospital billed for the identical code. The "high-volume" universe is every DRG reported by at least 100 hospitals over at least 1,000 national discharges — 219 of the 540 codes in the file — so each percentile rests on a large, stable set of hospitals rather than a handful of outliers.

The middle of that distribution is tight and high: the 25th-percentile DRG spread is 3.5×, the 75th-percentile DRG spread is 4.0×, and 210 of the 219 DRGs (96%) carry a spread of at least 3×. Wide hospital-to-hospital variation is not confined to exotic procedures or a few outlier facilities — it is the standing condition of almost every common code Medicare pays.

The DRG is fixed and so is what Medicare pays for it. What hospitals charge for it is not — for the same code, the dearest bill a median 3.8× the cheapest.

What it looks like on the most common codes

Take the codes Medicare pays most often, and the spread holds across all of them. Each row below is the same DRG, priced across thousands of hospitals: the 10th-percentile hospital's average charge, the median, the 90th, and the single highest.

Sepsis (DRG 871) is the clearest case because it is the most common: 2,661 hospitals bill it, and they bill it from $32,161 at the 10th percentile to $140,999 at the 90th — a $108,838 gap for the same coded admission, before the long tail up to a $591,962 high. The clinical content is, by definition of the code, the same severity tier of the same condition. The price is not.

The widest-variation codes

Rank the same 219 DRGs by spread instead of volume, and behavioral-health and complex-medical codes rise to the top. Psychoses lead: the 90th-percentile hospital billed 5.0× the 10th-percentile hospital for the identical code.

Charges, payments, and which gap is which

Two different gaps live in this file, and they are easy to confuse. The first is the charge-to-Medicare-payment gap — how far above Medicare's payment a hospital's charge sits — which runs 6.1× across all DRGs in 2024. The second, this study's subject, is the hospital-to-hospital spread on the same code. The charge-to-payment multiple itself varies by DRG, because hospitals mark up some codes more aggressively than others:

The charge figures here are discharge-weighted national averages. The Medicare payment is what actually changed hands. A stroke admission (DRG 065) carries a 7.5× charge-to-payment multiple; heart failure (DRG 291) a 5.6×. Neither multiple is what a patient or Medicare pays — Medicare pays the right-hand column — but together with the hospital-to-hospital spread above, they show charges behaving as a markup policy decoupled from both cost and reimbursement.

How to read these numbers

The percentiles and the payment column answer different questions, so read them in order:

1. The percentile columns (10th / median / 90th) are computed across hospitals for one DRG. Each hospital contributes one observation — its own average charge for that code — and a hospital that billed the code five times counts the same as one that billed it five thousand times. This is deliberate: the question is how hospitals differ in pricing, not how many patients each saw.
2. The 90/10 ratio is the spread in a single number. A ratio of 4.0× means the hospital at the 90th percentile of charges billed four times what the hospital at the 10th percentile billed for the identical code.
3. The charge-to-payment multiple is a different axis entirely — discharge-weighted, comparing total charges to total Medicare payments — and is included only to separate it from the spread, not as this study's finding.

A high spread on a code does not name any hospital as overpriced or underpriced. It says that the same federally defined unit of care carries radically different list prices depending only on where it is billed.

Methodology

Every figure is a direct aggregation over one public, row-level-signed Postgres table: inpatient_utilization_summary, built from the CMS Medicare Inpatient Hospitals by Provider and Service public-use file (MUP_IHP). The table carries one row per hospital (ccn) × MS-DRG (ms_drg_code) × program year (data_year); it is RLS Pattern B, public read. The snapshot used here was pulled 2026-06-11 and covers program year 2024, the most recent annual release.

The program year is resolved as max(data_year) at query time, never hard-coded, so the figures advance when CMS publishes the next file. Charge percentiles are computed with percentile_cont over avg_covered_charges, grouped by ms_drg_code, across the hospitals reporting that code — one hospital, one observation, unweighted, because the question is dispersion across hospitals rather than per-stay national average. The high-volume universe is restricted to DRGs with at least 100 reporting hospitals and at least 1,000 national discharges (219 of 540 codes) so each percentile is stable. The charge-to-Medicare multiple, by contrast, is discharge-weighted — sum(charges × discharges) / sum(payments × discharges) — to reproduce the national per-stay ratio.

CMS suppresses any hospital-DRG cell with fewer than 11 discharges; those cells arrive null and are excluded rather than imputed. The exact SQL is in the reproducibility block below, and the row-level signing contract is documented in the provenance methodology. Methodology version: drg-variation/v1.

Limitations

• Medicare fee-for-service only. The file covers traditional Medicare Part A inpatient stays. It excludes Medicare Advantage, Medicaid, and commercial admissions, so the spread describes Medicare-billed activity, not every payer.
• Charges are list prices, not transactions. Average covered charges are what hospitals billed, never what Medicare or a patient paid. The spread is a spread in markup policy, not in what care cost or what was collected.
• Within-DRG case mix is not fully controlled. A single MS-DRG fixes the severity tier, but residual differences in patient acuity, length of stay, and regional input costs contribute to the spread alongside chargemaster policy. The figures bound the variation; they do not attribute all of it to pricing discretion.
• Hospital-level, not stay-level percentiles. Each hospital contributes one averaged observation per DRG. The spread is across hospitals, so it understates the full range an individual patient might encounter.
• Suppressed small cells. Hospital-DRG combinations under 11 discharges are withheld by CMS and excluded, which slightly narrows the observed spread at low-volume hospitals.
• One program year, one snapshot. The figures are point-in-time: program year 2024, snapshot 2026-06-11. DRG definitions and payment weights are revised annually, so they are not directly comparable across years without re-basing.
• A pricing-transparency signal, not a quality signal. These figures describe how hospitals price a code, never how well any hospital performs. No facility is ranked, scored, or named in this study.

Sources

• CMS — Medicare Inpatient Hospitals by Provider and Service — the inpatient DRG public-use file (MUP_IHP), program year 2024, the sole source for this study.
• CMS — Medicare Inpatient Hospitals methodology — CMS's own definition of covered charges, payments, and the suppression rule for small cells.
• CMS — MS-DRG classifications and software — the official MS-DRG grouping logic and annual definitions manual.
• CMS — Hospital Price Transparency — the regulatory backdrop for hospital list-price (chargemaster) disclosure.

The companion dataset page for the CMS Medicare Inpatient Hospitals file lists the full schema and refresh cadence; the national per-code counterpart sits in the DRG cost reference, and broader hospital-finance context in hospitals running low on days cash on hand and who owns America's hospitals and which model makes money.