{
  "study": {
    "slug": "medicare-inpatient-drg-charge-variation-2026",
    "title": "Same DRG, wildly different price: hospital charge variation, 2024",
    "standfirst": "For the same Medicare DRG, hospitals bill wildly different amounts: across 219 high-volume codes in 2024, the 90th-percentile hospital charged a median of 3.8× what the 10th-percentile hospital charged for the identical stay. For sepsis — the most common code, billed by 2,661 hospitals — the spread is 4.4×.",
    "desk": "financial-distress",
    "article_type": "Original Research",
    "published": "2026-06-14",
    "issue": 67,
    "doi": "10.5072/fonteum/medicare-inpatient-drg-charge-variation-2026",
    "url": "https://fonteum.com/research/medicare-inpatient-drg-charge-variation-2026",
    "methodology_version": "drg-variation/v1"
  },
  "data_as_of": "2026-06-11",
  "datasets": [
    {
      "slug": "cms-inpatient-utilization",
      "name": "CMS Medicare Inpatient Hospitals",
      "publisher": "CMS — Medicare Inpatient Hospitals, by Provider and Service",
      "upstream_url": null
    }
  ],
  "key_findings": [
    {
      "number": "3.8×",
      "finding": "median ratio between the 90th- and 10th-percentile hospital's billed charge for the same MS-DRG, across 219 high-volume codes in 2024",
      "dataset": "cms-inpatient-utilization"
    },
    {
      "number": "4.4×",
      "finding": "the spread for sepsis (DRG 871), the most common Medicare code — $32,161 at the 10th-percentile hospital vs $140,999 at the 90th, for the identical stay",
      "dataset": "cms-inpatient-utilization"
    },
    {
      "number": "210 of 219",
      "finding": "high-volume DRGs where the 90th-percentile hospital billed at least 3× the 10th-percentile hospital for the same code — variation is the rule, not the exception",
      "dataset": "cms-inpatient-utilization"
    },
    {
      "number": "$108,838",
      "finding": "the difference in billed charge for one sepsis admission between a 10th-percentile and a 90th-percentile hospital — same DRG, same program, same year",
      "dataset": "cms-inpatient-utilization"
    }
  ],
  "faqs": [
    {
      "q": "Why do hospitals charge different amounts for the same DRG?",
      "a": "Billed charges are list prices each hospital sets on its own chargemaster — there is no national price for a DRG's charge. 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 charge reflects each hospital's markup policy, not the cost of care or what Medicare pays."
    },
    {
      "q": "Does the patient or Medicare pay these different charges?",
      "a": "No. Medicare pays its own administratively set DRG rate regardless of what a hospital charges, so the wide charge spread does not change Medicare's payment. Charges still matter: they anchor out-of-network and self-pay bills, and they are the list price an uninsured patient is first billed before any discount."
    },
    {
      "q": "Which DRG has the widest hospital-to-hospital charge spread?",
      "a": "Among high-volume codes, psychoses (DRG 885) has the widest spread: the 90th-percentile hospital billed 5.0× the 10th-percentile hospital ($89,309 vs $17,820), across 562 hospitals. Alcohol or drug dependence without rehabilitation therapy (DRG 897) is next at 4.9×. Behavioral-health codes cluster at the top of the variation ranking."
    },
    {
      "q": "How much does sepsis cost across hospitals?",
      "a": "For sepsis with a major complication (DRG 871) — the single most common Medicare inpatient code, billed by 2,661 hospitals in 2024 — the 10th-percentile hospital billed $32,161, the median $65,249, and the 90th-percentile $140,999. That is a 4.4× spread, or a $108,838 difference between the 10th- and 90th-percentile bill for the same admission."
    },
    {
      "q": "Is this the same as the charge-to-payment gap?",
      "a": "No — they are two different gaps. The charge-to-payment gap is how far one hospital's charge sits above Medicare's payment (6.1× on average across all DRGs in 2024). This study measures the hospital-to-hospital spread: how far hospitals differ from each other on the same code. We document the charge-to-payment side in the companion DRG cost reference."
    },
    {
      "q": "What program year and source does this cover?",
      "a": "Calendar year 2024, the most recent annual release of the CMS Medicare Inpatient Hospitals by Provider and Service public-use file (MUP_IHP), snapshotted 2026-06-11. The query resolves the latest available data_year at run time, so the figures advance automatically when CMS publishes the next annual file."
    },
    {
      "q": "Can I reproduce these figures?",
      "a": "Yes. Every number is a direct aggregation over the public inpatient_utilization_summary table — per-DRG percentiles of billed charges across hospitals, for the most recent program year. The exact SQL is published in the reproducibility block below; a re-run resolves to the same rows in the frozen 2026-06-11 snapshot."
    }
  ],
  "citation": {
    "apa": "Fonteum Research. (2026, June 14). Same DRG, wildly different price: hospital charge variation, 2024. Fonteum Research, Issue 67. https://doi.org/10.5072/fonteum/medicare-inpatient-drg-charge-variation-2026",
    "url": "https://fonteum.com/research/medicare-inpatient-drg-charge-variation-2026"
  },
  "reproducible_sql": "-- Medicare inpatient DRG charge VARIATION — fully reproducible query.\n--\n-- Question: for the SAME MS-DRG, how widely do hospitals differ in what they\n-- bill? We measure the hospital-to-hospital spread of average covered (billed)\n-- charges per DRG — the 10th / median / 90th percentile across hospitals and\n-- the 90/10 ratio — plus the per-DRG charge-to-Medicare-payment multiple.\n-- This is DISTINCT from the charge-to-payment gap of the DRG cost reference:\n-- here the axis is hospital-vs-hospital on one code, not charge-vs-payment.\n--\n-- Source:\n--   public.inpatient_utilization_summary — CMS \"Medicare Inpatient Hospitals,\n--     by Provider and Service\" public-use file (MUP_IHP). One row per\n--     hospital (CCN) × MS-DRG × data_year. public, read-only.\n--     Snapshot 2026-06-11; program year 2024 (the most recent annual release).\n--     145,879 rows · 2,906 hospitals · 540 distinct MS-DRGs · 4,952,481 discharges.\n--     License: US-Government-Works (17 U.S.C. §105).\n--\n-- Grain note: CMS suppresses any hospital-DRG cell with fewer than 11\n-- discharges (total_discharges IS NULL). Those cells are excluded — never\n-- imputed. avg_covered_charges is a per-hospital average for that DRG.\n--\n-- Percentiles are computed ACROSS HOSPITALS, one hospital = one observation\n-- (unweighted), because the question is dispersion across hospitals, not the\n-- per-stay national average. The charge-to-payment multiple, by contrast, is\n-- discharge-weighted so it reproduces the national per-stay ratio.\n--\n-- High-volume universe: DRGs reported by >=100 hospitals over >=1,000 national\n-- discharges (219 of the 540 codes) so each percentile is stable.\n--\n-- \"Most recent program year\" is resolved at query time, never hard-coded.\n\nWITH latest AS (\n  SELECT max(data_year) AS yr FROM public.inpatient_utilization_summary\n),\nbase AS (\n  SELECT ccn,\n         ms_drg_code,\n         ms_drg_description,\n         total_discharges,\n         avg_covered_charges::numeric    AS cc,   -- per-hospital avg billed charge\n         avg_medicare_payments::numeric  AS mp    -- per-hospital avg Medicare payment\n  FROM public.inpatient_utilization_summary\n  WHERE data_year = (SELECT yr FROM latest)\n    AND total_discharges IS NOT NULL              -- drop CMS-suppressed cells\n    AND avg_covered_charges IS NOT NULL\n),\nper_drg AS (\n  SELECT\n    ms_drg_code,\n    max(ms_drg_description)                                        AS descr,\n    count(DISTINCT ccn)                                            AS hospitals,\n    sum(total_discharges)                                          AS discharges,\n    round(percentile_cont(0.1) WITHIN GROUP (ORDER BY cc)::numeric) AS p10,\n    round(percentile_cont(0.5) WITHIN GROUP (ORDER BY cc)::numeric) AS median_charge,\n    round(percentile_cont(0.9) WITHIN GROUP (ORDER BY cc)::numeric) AS p90,\n    round(max(cc))                                                 AS max_charge,\n    round((percentile_cont(0.9) WITHIN GROUP (ORDER BY cc)\n          / nullif(percentile_cont(0.1) WITHIN GROUP (ORDER BY cc), 0))::numeric, 1)\n                                                                   AS p90_p10_ratio,\n    round(sum(cc * total_discharges) / sum(total_discharges))      AS dw_charge,\n    round(sum(mp * total_discharges) / sum(total_discharges))      AS dw_pay,\n    round((sum(cc * total_discharges)\n          / nullif(sum(mp * total_discharges), 0))::numeric, 1)    AS charge_to_pay\n  FROM base\n  GROUP BY ms_drg_code\n),\nhi_vol AS (                                  -- the 219-DRG high-volume universe\n  SELECT * FROM per_drg WHERE hospitals >= 100 AND discharges >= 1000\n)\n\n-- ============================================================================\n-- (1) Headline: how wide is the spread across the high-volume universe?\n--     The median DRG's 90/10 charge ratio is the lead figure (3.78×).\n-- ============================================================================\nSELECT\n  count(*)                                                              AS drgs_qualifying,\n  round(percentile_cont(0.25) WITHIN GROUP (ORDER BY p90_p10_ratio), 2) AS p25_ratio,\n  round(percentile_cont(0.5)  WITHIN GROUP (ORDER BY p90_p10_ratio), 2) AS median_ratio,\n  round(percentile_cont(0.75) WITHIN GROUP (ORDER BY p90_p10_ratio), 2) AS p75_ratio,\n  min(p90_p10_ratio)                                                    AS min_ratio,\n  max(p90_p10_ratio)                                                    AS max_ratio,\n  count(*) FILTER (WHERE p90_p10_ratio >= 3)                            AS drgs_ge_3x,\n  count(*) FILTER (WHERE p90_p10_ratio >= 4)                            AS drgs_ge_4x\nFROM hi_vol;\n--  drgs_qualifying p25  median p75  min  max  ge_3x ge_4x\n--  219             3.52 3.78   4.02 2.62 5.01 210   58\n\n-- ============================================================================\n-- (2) The spread on the most COMMON codes — same DRG, priced across hospitals.\n--     (Ranked by discharge volume; the study's first table.)\n-- ============================================================================\nSELECT ms_drg_code, descr, hospitals, p10, median_charge, p90, max_charge, p90_p10_ratio\nFROM hi_vol\nWHERE ms_drg_code IN ('871','291','177','193','683','065','470','190')\nORDER BY discharges DESC;\n--  871 Sepsis w/ MCC               2,661  32,161  65,249 140,999 591,962 4.4\n--  291 Heart failure w/ MCC        2,587  22,190  43,409  96,115 437,712 4.3\n--  177 Respiratory infections      2,332  30,023  55,022 118,145 487,216 3.9\n--  193 Simple pneumonia w/ MCC     2,442  24,559  46,587 101,466 375,925 4.1\n--  683 Renal failure w/ CC         2,018  18,296  32,957  71,470 282,993 3.9\n--  065 Intracranial hemorrhage     1,727  25,540  45,544 100,197 363,815 3.9\n--  470 Hip/knee replacement        1,212  42,710  79,947 167,283 383,606 3.9\n--  190 COPD w/ MCC                 1,595  22,754  41,850  86,174 314,888 3.8\n\n-- ============================================================================\n-- (3) The WIDEST-variation codes — same universe, ranked by 90/10 ratio.\n-- ============================================================================\nSELECT ms_drg_code, descr, hospitals, p10, median_charge, p90, p90_p10_ratio\nFROM hi_vol\nORDER BY p90_p10_ratio DESC\nLIMIT 10;\n--  885 Psychoses                              562  17,820  36,077  89,309 5.0\n--  897 Alcohol/drug abuse w/o rehab, w/o MCC  502  15,561  32,548  75,625 4.9\n--  565 Other musculoskeletal dx w/ CC         119  19,575  42,155  91,081 4.7\n--  542 Pathological fractures w/ MCC          184  38,338  68,552 177,576 4.6\n--  196 Interstitial lung disease w/ MCC       217  35,405  80,187 161,953 4.6\n--  809 Major hematological/immuno dx          203  27,887  55,001 128,304 4.6\n--  554 Bone diseases/arthropathies w/o MCC    200  14,430  32,190  64,800 4.5\n--  884 Organic disturbances/intellectual dis  707  25,102  49,954 112,207 4.5\n--  103 Headaches w/o MCC                      110  22,920  46,101 100,811 4.4\n--  871 Sepsis w/ MCC                        2,661  32,161  65,249 140,999 4.4\n\n-- ============================================================================\n-- (4) The OTHER gap — per-DRG charge-to-Medicare-payment multiple (a different\n--     axis: charge-vs-payment, discharge-weighted). Included to separate it\n--     from the hospital-to-hospital spread above.\n-- ============================================================================\nSELECT ms_drg_code, descr, dw_charge, dw_pay, charge_to_pay\nFROM per_drg\nWHERE ms_drg_code IN ('065','392','312','470','683','193','871','291')\nORDER BY charge_to_pay DESC;\n--  065 Intracranial hemorrhage       56,315   7,474  7.5\n--  392 Esophagitis/gastroenteritis   40,642   5,662  7.2\n--  312 Syncope and collapse          46,199   6,636  7.0\n--  470 Hip/knee replacement          92,976  14,289  6.5\n--  683 Renal failure w/ CC           41,156   6,630  6.2\n--  193 Simple pneumonia w/ MCC       60,101   9,868  6.1\n--  871 Sepsis w/ MCC                 90,381  15,524  5.8\n--  291 Heart failure w/ MCC          56,495  10,022  5.6\n\n-- Overall, all DRGs: discharge-weighted charge $92,408 vs payment $15,166 =\n-- 6.09× charge-to-Medicare multiple (the all-file figure of the cost reference).",
  "license": "U.S. Government Works (federal sources; 17 U.S.C. §105)",
  "generated_by": "Fonteum — https://fonteum.com",
  "notes": "Aggregate, source-traced figures frozen to the snapshot above. Reproduce by running reproducible_sql against the cited federal dataset; no per-entity records are included."
}