cms-inpatient-utilization · CMS
cms-inpatient-utilization · CMS
cms-inpatient-utilization · CMS
cms-inpatient-utilization · CMS
A DRG code is how Medicare buys a hospital stay. Rather than itemize every test, drug, and bed-day, the program assigns each inpatient admission to one MS-DRG — a Medicare Severity Diagnosis-Related Group — based on the patient's diagnoses and procedures, and pays a single bundled rate for that group. The code is the price tag. Knowing what each one pays, and how often it is billed, is the entry point to almost any question about Medicare hospital spending.
This is a reference. It aggregates the CMS Medicare Inpatient Hospitals by Provider and Service public-use file — one row per hospital per DRG — up to the national level, for every MS-DRG present in the most recent program year. For each code it reports four numbers: how many times Medicare paid it (discharges), what hospitals billed for it (average covered charges), what Medicare actually paid (average Medicare payment), and how many hospitals reported it (provider count).
What the 2024 file shows
The 2024 release covers 540 distinct MS-DRGs, billed across 2,906 hospitals over 4,952,481 discharges. Discharge-weighted across all of them, Medicare paid an average of $15,166 per stay — against $92,408 in average billed charges. The gap between what hospitals charge and what Medicare pays is 6.1×, and it holds, with wide variation, across nearly every code in the file.
Two facts make the reference useful at a glance. The volume is concentrated: the busiest handful of DRGs account for a large share of all discharges, and they are overwhelmingly medical — sepsis, heart failure, pneumonia, kidney infection. The cost is concentrated somewhere else entirely: the most expensive DRGs are rare surgical and cellular therapies that almost no patient receives. A DRG that tops one ranking rarely appears on the other.
A DRG code is a price the government sets, not the hospital. Across 540 of them, Medicare paid roughly a sixth of what hospitals charged.
The highest-volume DRGs
Ranked by national discharges, the top of the 2024 reference is a list of common, acute medical admissions. Sepsis without prolonged ventilation, with a major complication or comorbidity (MCC), is the single most-billed Medicare DRG in the country.
| MS-DRG | Description | Hospitals | Discharges | Avg covered charge | Avg Medicare payment |
|---|---|---|---|---|---|
| 871 | Sepsis w/o prolonged ventilation, w/ MCC | 2,661 | 577,119 | $90,381 | $15,524 |
| 291 | Heart failure and shock, w/ MCC | 2,587 | 304,694 | $56,495 | $10,022 |
| 177 | Respiratory infections and inflammations, w/ MCC | 2,332 | 141,351 | $71,065 | $12,668 |
| 193 | Simple pneumonia and pleurisy, w/ MCC | 2,442 | 136,649 | $60,101 | $9,868 |
| 872 | Sepsis w/o prolonged ventilation, w/o MCC | 2,216 | 103,998 | $49,416 | $7,681 |
| 690 | Kidney and urinary tract infections, w/o MCC | 2,133 | 91,493 | $38,134 | $5,986 |
| 189 | Pulmonary edema and respiratory failure | 2,098 | 89,991 | $57,943 | $9,974 |
| 280 | Acute myocardial infarction, discharged alive, w/ MCC | 1,825 | 83,841 | $76,887 | $12,400 |
The per-stay Medicare payments here cluster between roughly $6,000 and $16,000. None of them is individually expensive; they matter because they recur hundreds of thousands of times. The two sepsis codes alone — DRG 871 and DRG 872 — account for more than 680,000 discharges, the largest single clinical category in Medicare inpatient care.
The costliest DRGs
Rank the same reference by Medicare payment instead, holding to DRGs with at least 1,000 national discharges so the top is a stable reference rather than a handful of cases, and the list changes completely. These are transplants, cellular therapies, and extended critical care.
| MS-DRG | Description | Hospitals | Discharges | Avg covered charge | Avg Medicare payment |
|---|---|---|---|---|---|
| 018 | CAR T-cell and other immunotherapies | 49 | 1,360 | $2,187,781 | $434,771 |
| 001 | Heart transplant or implant of heart assist system, w/ MCC | 87 | 1,814 | $1,604,490 | $295,327 |
| 003 | ECMO or tracheostomy w/ prolonged ventilation | 272 | 6,749 | $1,160,394 | $209,580 |
| 004 | Tracheostomy w/ prolonged ventilation, w/o major O.R. procedure | 212 | 3,589 | $680,475 | $120,704 |
| 005 | Liver transplant w/ MCC, or intestinal transplant | 53 | 1,043 | $819,999 | $116,001 |
| 453 | Combined anterior and posterior spinal fusion, w/ MCC | 57 | 1,012 | $588,627 | $91,243 |
| 216 | Cardiac valve procedures w/ cardiac catheterization, w/ MCC | 124 | 2,577 | $501,084 | $90,313 |
| 219 | Cardiac valve procedures w/o cardiac catheterization | 350 | 12,333 | $393,302 | $69,387 |
The costliest single code, CAR T-cell immunotherapy (DRG 018), averages $434,771 per stay — 29 times the all-DRG average. It was billed by just 49 hospitals across 1,360 discharges. The provider-count column is the tell: high-cost DRGs are performed at a narrow set of academic and specialty centers, while the high-volume DRGs above are billed by more than two thousand hospitals each.
A few DRGs sit in both economies. DRG 853 — infectious and parasitic diseases with operating-room procedures, with MCC — drew 79,560 discharges (a top-15 volume) at an average Medicare payment of $40,557 (well above the typical medical DRG). It is the rare code that is both common and expensive, and it is where high-acuity infection care and surgery overlap.Charges, payments, and the 6.1× gap
The two dollar columns in the reference measure different things, and the distinction is the most common source of confusion about hospital prices. Average covered charges are the hospital's list price — the amount it billed. Average Medicare payment is what Medicare actually paid under the DRG rate. Charges are set by the hospital and are rarely paid in full by anyone; the DRG payment is administratively set and is what changes hands.
Across the whole 2024 file, charges run 6.1× the Medicare payment — $92,408 against $15,166. The multiple is not uniform: it is widest on the high-charge surgical DRGs (CAR T-cell therapy billed $2.19M against a $435K payment, roughly 5×) and on common medical DRGs alike. We documented the facility-level version of this charge-to-payment spread, and how it varies hospital by hospital, in the hospital charges by DRG study. The reference here is the national counterpart: one row per DRG, every DRG in the file.
How to read a DRG code
MS-DRGs are not arbitrary. The numbering follows the body system and the kind of care, and most codes come in graded pairs or triples that encode severity:
- The base group is a clinical category — sepsis, heart failure, joint replacement, a transplant.
- The severity tier is appended as a suffix in the description: with MCC (major complication or comorbidity), with CC (complication or comorbidity), or without CC/MCC. Each tier up pays more, because sicker patients cost more to care for.
- The payment follows the tier. DRG 871 (sepsis with MCC) pays $15,524; its sibling DRG 872 (sepsis without MCC) pays $7,681 — the same condition, half the rate, decided by the complication coding.
This is why two codes that look almost identical can sit far apart in the reference, and why the with MCC variants dominate the volume rankings: the sickest version of each common condition is both the most expensive to treat and, among Medicare's older and frailer population, often the most frequently billed.
Across all 540 DRGs, the median code is billed by 53 hospitals and carries a median Medicare payment of $15,170 — almost exactly the discharge-weighted national average, a coincidence that reflects how the common medical DRGs anchor the middle of the distribution. 260 of the 540 DRGs clear 1,000 national discharges; the long tail below that is the rare-procedure and rare-diagnosis codes.
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 reference advances when CMS publishes the next file. Per-DRG aggregates are computed by GROUP BY ms_drg_code over the hospital-level rows. Dollar averages are discharge-weighted — sum(metric × total_discharges) / sum(total_discharges) — so a small hospital does not count the same as a large one, and the result reproduces the per-stay average a payer would observe nationally. Provider counts are count(DISTINCT ccn) per DRG. The charge-to-Medicare multiple is the ratio of total billed charges to total Medicare payments, both discharge-weighted.
CMS suppresses any hospital-DRG cell with fewer than 11 discharges; those cells arrive null and are excluded from every weighted average 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-reference/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 volumes are a floor on total US inpatient activity, not the whole.
- Charges are list prices, not transactions. Average covered charges are what hospitals billed, not what anyone paid. They are useful as a relative signal across DRGs, not as a measure of cost.
- Suppressed small cells. Hospital-DRG combinations under 11 discharges are withheld by CMS and excluded here. For rare DRGs at low-volume hospitals, the national totals understate true activity by an unknown but small margin.
- One program year, one snapshot. The reference is point-in-time: program year 2024, snapshot 2026-06-11. DRG definitions and payment weights are revised annually, so figures are not directly comparable across years without re-basing.
- Descriptions are abbreviated. MS-DRG titles in the tables are shortened for readability; the code is the authoritative key, and the full CMS long title governs.
- A payment reference, not a quality signal. These figures describe what Medicare pays and how often, never how well any hospital performs. No facility is ranked, scored, or surfaced by name 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 reference.
- CMS — Medicare Inpatient Hospitals methodology — CMS's own definition of 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.
- 42 CFR Part 412 — Prospective Payment Systems for Inpatient Hospital Services — the regulatory basis for DRG-based inpatient payment.
The companion dataset page for the CMS Medicare Inpatient Hospitals file lists the full schema and refresh cadence; broader hospital-finance context sits in hospitals running low on days cash on hand and who owns America's hospitals and which model makes money.
Frequently asked questions
- What is a DRG code?
- A DRG — diagnosis-related group — is the unit Medicare pays hospitals by. Each inpatient stay is assigned one MS-DRG (Medicare Severity DRG) from its diagnoses and procedures, and Medicare pays a set rate for that group rather than itemizing the bill. There are 540 MS-DRGs with Medicare inpatient activity in the 2024 file.
- What is the average Medicare payment per inpatient stay?
- $15,166 across all DRGs in 2024, discharge-weighted across 2,906 hospitals and 4,952,481 discharges. The figure varies enormously by DRG: a urinary-tract-infection stay (DRG 690) averages $5,986, while a CAR T-cell immunotherapy admission (DRG 018) averages $434,771.
- Which DRG has the most Medicare discharges?
- DRG 871 — septicemia or severe sepsis without prolonged mechanical ventilation, with major complication — at 577,119 discharges in 2024, billed by 2,661 hospitals. It is the single most common Medicare inpatient DRG, ahead of heart failure (DRG 291) at 304,694 discharges.
- Why are hospital charges so much higher than Medicare payments?
- Billed charges are list prices set by each hospital and rarely paid in full by anyone. Medicare pays its own administratively set DRG rate regardless of charges. Across all DRGs in 2024, average charges were $92,408 and the average Medicare payment was $15,166 — a 6.1× gap.
- Are these the prices a patient pays?
- No. These are the amounts Medicare pays the hospital for the stay, plus the hospital's billed charges. A Medicare beneficiary's own cost is set separately by deductibles and coinsurance, not by the DRG payment shown here. The figures are facility-side reimbursement, not patient out-of-pocket cost.
- What program year does this reference cover?
- Calendar year 2024, the most recent annual release of the CMS Medicare Inpatient Hospitals by Provider and Service public-use file, snapshotted 2026-06-11. The query resolves the latest available data_year at run time, so the reference advances automatically when CMS publishes the next annual file.
- Can I reproduce these figures?
- Yes. Every number is a direct aggregation over the public inpatient_utilization_summary table, grouped by MS-DRG for the most recent program year. The exact SQL is published in the reproducibility block below; the page, the figures, and a re-run of the query all resolve to the same rows in the frozen 2026-06-11 snapshot.
Datasets used
Reproducibility
Every claim, reproducible
The SQL
-- Medicare inpatient DRG cost-and-volume reference — fully reproducible query.
--
-- Question: across every MS-DRG that Medicare paid for inpatient care in the
-- most recent program year, what did each code cost and how often was it billed?
-- For each DRG: total discharges, average covered (billed) charges, average
-- Medicare payment, and the number of distinct hospitals reporting it — ranked.
--
-- Source:
-- public.inpatient_utilization_summary — CMS "Medicare Inpatient Hospitals,
-- by Provider and Service" public-use file (MUP_IHP). One row per
-- hospital (CCN) × MS-DRG × data_year. RLS Pattern B — public read.
-- Snapshot 2026-06-11; program year 2024 (the most recent annual release).
-- 145,879 rows · 2,906 hospitals · 540 distinct MS-DRGs · 4,952,481 discharges.
-- License: US-Government-Works (17 U.S.C. §105).
--
-- Grain note: CMS suppresses any hospital-DRG cell with fewer than 11
-- discharges (total_discharges IS NULL). Those cells are excluded from
-- discharge-weighted averages below — never imputed.
--
-- Averages are DISCHARGE-WEIGHTED across hospitals, not a mean-of-means: each
-- hospital's per-stay figure is weighted by that hospital's discharge count, so
-- a 5-discharge hospital does not count the same as a 5,000-discharge one. This
-- reproduces the national per-stay average a payer would actually observe.
--
-- "Most recent program year" is resolved at query time, never hard-coded.
WITH latest AS (
SELECT max(data_year) AS yr FROM public.inpatient_utilization_summary
),
base AS (
SELECT *
FROM public.inpatient_utilization_summary
WHERE data_year = (SELECT yr FROM latest)
AND total_discharges IS NOT NULL -- drop CMS-suppressed cells
)
-- ============================================================================
-- (1) Headline overview for the most recent program year.
-- ============================================================================
SELECT
(SELECT yr FROM latest) AS data_year,
count(*) AS hospital_drg_rows,
count(DISTINCT ccn) AS hospitals,
count(DISTINCT ms_drg_code) AS drgs,
sum(total_discharges) AS total_discharges,
round(sum(avg_covered_charges * total_discharges) / sum(total_discharges)) AS dw_avg_covered_charge,
round(sum(avg_medicare_payments * total_discharges) / sum(total_discharges)) AS dw_avg_medicare_payment,
round(sum(avg_total_payments * total_discharges) / sum(total_discharges)) AS dw_avg_total_payment,
round(
sum(avg_covered_charges * total_discharges)
/ nullif(sum(avg_medicare_payments * total_discharges), 0), 2
) AS charge_to_medicare_multiple
FROM base;
-- data_year hospital_drg_rows hospitals drgs total_discharges dw_avg_covered_charge dw_avg_medicare_payment dw_avg_total_payment charge_to_medicare_multiple
-- 2024 145,879 2,906 540 4,952,481 92,408 15,166 18,360 6.09
-- ============================================================================
-- (2) THE REFERENCE — per-DRG aggregates, ranked by discharge volume.
-- This is the canonical table: every DRG present in the program year, with
-- discharge volume, billed charges, Medicare payment, and provider count.
-- Swap the ORDER BY to rank by cost (dw_avg_medicare_payment DESC) instead.
-- ============================================================================
SELECT
ms_drg_code,
max(ms_drg_description) AS ms_drg_description,
count(DISTINCT ccn) AS hospitals,
sum(total_discharges) AS discharges,
round(sum(avg_covered_charges * total_discharges) / sum(total_discharges)) AS avg_covered_charge,
round(sum(avg_medicare_payments * total_discharges) / sum(total_discharges)) AS avg_medicare_payment,
round(sum(avg_total_payments * total_discharges) / sum(total_discharges)) AS avg_total_payment
FROM base
GROUP BY ms_drg_code
ORDER BY discharges DESC; -- full set: 540 rows
-- Highest-volume rows (top of the 540-row reference):
-- ms_drg_code description hospitals discharges avg_covered_charge avg_medicare_payment
-- 871 Septicemia/severe sepsis w/o MV >96 hrs, w/ MCC 2,661 577,119 90,381 15,524
-- 291 Heart failure and shock, w/ MCC 2,587 304,694 56,495 10,022
-- 177 Respiratory infections and inflammations, w/ MCC 2,332 141,351 71,065 12,668
-- 193 Simple pneumonia and pleurisy, w/ MCC 2,442 136,649 60,101 9,868
-- 872 Septicemia/severe sepsis w/o MV >96 hrs, w/o MCC 2,216 103,998 49,416 7,681
-- 690 Kidney and urinary tract infections, w/o MCC 2,133 91,493 38,134 5,986
-- ============================================================================
-- (3) Same reference, ranked by cost (Medicare payment per stay), volume floor
-- of 1,000 national discharges so the top is a stable reference, not a
-- handful-of-cases artifact. 260 of the 540 DRGs clear the 1,000 floor.
-- ============================================================================
SELECT
ms_drg_code,
max(ms_drg_description) AS ms_drg_description,
count(DISTINCT ccn) AS hospitals,
sum(total_discharges) AS discharges,
round(sum(avg_covered_charges * total_discharges) / sum(total_discharges)) AS avg_covered_charge,
round(sum(avg_medicare_payments * total_discharges) / sum(total_discharges)) AS avg_medicare_payment
FROM base
GROUP BY ms_drg_code
HAVING sum(total_discharges) >= 1000
ORDER BY avg_medicare_payment DESC
LIMIT 10;
-- ms_drg_code description hospitals discharges avg_covered_charge avg_medicare_payment
-- 018 CAR T-cell and other immunotherapies 49 1,360 2,187,781 434,771
-- 001 Heart transplant or implant of heart assist system, w/ MCC 87 1,814 1,604,490 295,327
-- 003 ECMO or tracheostomy w/ MV >96 hrs (except face/mouth/neck) 272 6,749 1,160,394 209,580
-- 004 Tracheostomy w/ MV >96 hrs, w/o major O.R. procedure 212 3,589 680,475 120,704
-- 005 Liver transplant w/ MCC or intestinal transplant 53 1,043 819,999 116,001
-- ============================================================================
-- (4) Distribution context: how many DRGs clear a meaningful volume floor, and
-- the median provider count + median Medicare payment across all DRGs.
-- ============================================================================
WITH per_drg AS (
SELECT
ms_drg_code,
count(DISTINCT ccn) AS hospitals,
sum(total_discharges) AS discharges,
sum(avg_medicare_payments * total_discharges) / sum(total_discharges) AS avg_medicare_payment
FROM base
GROUP BY ms_drg_code
)
SELECT
count(*) AS drgs,
count(*) FILTER (WHERE discharges >= 1000) AS drgs_ge_1000_discharges,
round(percentile_cont(0.5) WITHIN GROUP (ORDER BY hospitals)) AS median_hospitals_per_drg,
round(percentile_cont(0.5) WITHIN GROUP (ORDER BY avg_medicare_payment))::int AS median_drg_medicare_payment
FROM per_drg;
-- drgs drgs_ge_1000_discharges median_hospitals_per_drg median_drg_medicare_payment
-- 540 260 53 15,170The snapshot
| dataset_id | cms-inpatient-utilization |
| snapshot_date | 2026-06-11 |
| sha256 | |
| doi | 10.5072/fonteum/medicare-inpatient-drg-cost-reference-2026 |
| slsa_provenance_url |
The JOINs
program year: data_year = max(data_year) in inpatient_utilization_summary -- resolved at query time, = 2024 matchable cells: total_discharges IS NOT NULL -- CMS suppresses any hospital-DRG cell under 11 discharges; never imputed per-DRG aggregate: GROUP BY ms_drg_code over hospital × DRG rows averages are discharge-weighted: sum(metric * total_discharges) / sum(total_discharges) -- not a mean-of-means provider count: count(DISTINCT ccn) per ms_drg_code charge-to-Medicare multiple = sum(charges*disch) / sum(medicare*disch) = 6.09
The pipeline version
| git_sha | |
| slsa_provenance | |
| methodology_version | drg-reference/v1 |
Reproduce this
Run the exact query against the frozen 2026-06-11.
Cite this study
Citation-ready for researchers and AI.
Check the chain
Each figure is snapshot-attested — re-derive the hash from the federal file.
cms-inpatient-utilization · 2026-06-11SHA-256 a3f1c9…7e6b- FINANCIAL DISTRESS · JUN 2026Hospitals running out of cash: the days-cash signal, and why most of it is a reporting artifactFederal HCRIS cost reports let us compute days cash on hand for 5,459 hospitals, but facility-level figures are distorted by system-level cash pooling — so the raw '2,800 hospitals under 30 days' headline is mostly noise. The defensible signal is narrower: 690 hospitals that report thin cash and also run an operating loss.
- FINANCIAL DISTRESS · JUN 2026Rural hospital closures, by the numbers: which hospitals are most at riskRural Critical Access Hospitals — the small facilities at the center of the closure crisis — run a 50.4% financial-distress rate, against 39.2% for urban hospitals, across 6,019 Medicare hospitals in the federal HCRIS cost reports. Their average operating margin is −8.93%, and 682 are losing money on patient care.
- FINANCIAL DISTRESS · JUN 2026For-profit, nonprofit, or government: who owns America's hospitals, and which model makes moneyAcross 6,019 US hospitals in the federal HCRIS cost reports, for-profit facilities are the only ownership class earning a positive average operating margin — +0.19% — while nonprofit hospitals average −4.75% and government hospitals −62.38%. The ranking holds on every measure, but the gap is narrower than the averages suggest.
- FINANCIAL DISTRESS · JUN 2026Hospital charity care, by the numbers: who actually gives the most free careNonprofit hospitals — tax-exempt in exchange for community benefit — deliver charity care worth just 1.53% of their patient revenue, the lowest share of any ownership type, below for-profit hospitals (3.00%) and less than half the government rate (3.76%), across $27.68 billion in free care in the federal HCRIS cost reports.
- FINANCIAL DISTRESS · JUN 2026The most expensive Medicare Part D drugs are rarely the most prescribedEliquis cost Medicare Part D $19.88 billion in 2024 — the single costliest drug in the program, yet only its 12th most-prescribed. That inversion defines Part D: brand-name drugs are 23.9% of prescriptions but 90.1% of the dollars, while cheap generics carry the volume and almost none of the cost.
Federal source citations
Fonteum Research · June 14, 2026 · All figures trace to the frozen federal-data snapshot cited above.