RAND Hospital Data - onetomapanalytics/Meta_Data GitHub Wiki

RAND Hospital Data

General description

  1. Database primary purpose - The RAND Hospital Data tool is an effort to enhance CMS HCRIS data to make them more accessible to a broad audience of academics, analysts, and hospital executives and their consultants who regularly work with the data to solve various problems and answer complex questions.
  2. Overall data type - Health facilities, hospital expenditures
  3. Dataset type - Longitudinal
  4. Data source - Claims
  5. Data level - Hospital level
  6. Geographic location of the data collection sites - United States
  7. Sponsor, manager, or home institution - RAND corporation
  8. Date range - 2017 - 2022
  9. Geolocation data - Hospital address, city, county, state, FIPS, ZIP code
  10. Dates - Cost reporting period beginning and end date, hospital and Federal fiscal year, calendar year
  11. Hospital identifiers - Name, Medicare number
  12. Longitudinal tracking - Track providers (see identifiers above)
  13. Financial variables - Cost reports, total unreimbursed and uncompensated care cost, administrative costs, cash flow margin, cost of charity care, costs per discharge, total salaries and employee benefits, estimated charges, revenues, expenses, and income
  14. Clinical areas of interest - all
  15. Number of records - 4,624 hospitals (May 2022)
  16. Variables that are uniquely present in this dataset - Enhanced CMS HCRIS data

Applicable methods

  1. Association methods, such as multivariate regression (1), generalized linear models (GLM) (2, 3), t-tests (4)
  2. Propensity scores (5)
  3. Causation, such as difference-in-differences (6, 7)
  4. Descriptive analyses (8)
  5. Time to event (9)

Data dictionary

To access the RAND data dictionary, click here

Variable categories

  1. Hospital characteristics (i.e., name, address, category of provider/supplier, rural/urban, beds, employees on payroll, ownership)
  2. Cost reports (i.e., period, status, record number, source)
  3. Financial variables (i.e., occupancy, operating margin, total unreimbursed and uncompensated care cost, charity care, cost per discharge, payer mix)

Linkage to other datasets

  • Linkages can be established for any dataset that might have geolocation data (i.e., FIPS or ZIP code) or hospital identifier (i.e., name or Medicare ID)