Peer distribution
Occupancy
Percentile describes position, not quality or efficiency. Each metric uses only peers with a valid value.
CMS cost reports · Hospital pilot
Hospital Benchmarking
Understand a hospital’s financial position and compare it with facilities that deliver similar care.
Loading CMS hospital reports…
Your comparison group
Define this hospital’s peers
Combine geography, category, operating type and size. The target hospital is always excluded.
Inspect and adjust peer membership
| Hospital / CCN | Location | Beds | Category / ownership | Membership |
|---|
Target versus your defined peers
How does this hospital compare?
All cards, distributions, history comparisons and exports use this peer group. Eligible peer reports span 300–430 days.
Longitudinal benchmarking
How is this hospital changing over time?
Follow the target, the peer median, and the middle 50% of peers. The main dataset selection defines the anchor hospital and peer group.
Copper: target · Dark teal: peer median · Shaded area: peer 25th–75th percentiles. Gaps indicate missing or ineligible reports, not zero. Each benchmark requires five valid peers. Annual change is shown only for adjacent datasets with eligible, nonoverlapping target periods separated by no more than 60 days. Percentiles describe position, not quality.
Volume and dollar flows are normalized to 365.25 days. Dollars are nominal and are not adjusted for inflation. Dataset labels do not standardize fiscal periods to calendar years; many reports end in the following calendar year. Growth rates require a positive baseline. Fixed membership does not guarantee unchanged services or ownership; missing reports do not establish closure.
Next data layers
System relationships & commercial prices
The reviewed RAND codebook supports system IDs, home-office links and hospital-based providers. Those observations are not loaded in this pilot. RAND’s public price-study annex is a separate claims-based source; price values and ownership relationships will appear after matching and validation.
Data & definitions
What these comparisons mean
Source and report selection
CMS Hospital Provider Cost Report records for 2017–2024 are retrieved through paginated JSON API queries. Raw CSV and ZIP distributions are not downloaded. The page loads a compact derived cache for fast static hosting. The source contains selected HCRIS measures, not every worksheet field. Where multiple reports exist for a CCN in a source dataset, the pilot prefers a 300–430-day report, then longest duration, latest end date and highest report ID. It does not add partial reports together.
Scopes, missing values and peer groups
Cost-report totals can include hospital-complex components. The initial peer group matches the target’s CCN category, clinical provider type and bed-size band. Users can redefine geography, category, ownership, clinical type, rural status, bed bands and minimum/maximum beds, or exclude individual peers; comparisons are descriptive and are not adjusted for case mix, teaching or local wage levels. Missing inputs and nonpositive denominators produce unavailable metrics. Occupancy outside 0–100% is excluded. No missing amount is treated as zero. Extreme remaining values are retained, with robust peer medians.
Bed bands: under 25, 25–99, 100–299, and 300+. Short or unusually long selected reports can be inspected but receive no peer percentile. Peer groups below five facilities are labeled too small for benchmarking. Percentiles use a midpoint rank for ties.
Metric definitions and RAND distinction
Patient-service margin is distinct from RAND operating margin. Total margin follows RAND’s documented net-income / (net-patient-revenue + other-income) formula using CMS fields. RAND’s exclusions for operating margin, CMI adjustments, prices and other supplemental measures cannot be reconstructed from this subset without further inputs.
Source: Centers for Medicare & Medicaid Services. Independent derived analysis; not affiliated with CMS or RAND. This pilot measures reported economics, not clinical quality.